admin/dify
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

chore(api/core): apply ruff reformatting (#7624)

Browse Source
This commit is contained in:
Bowen Liang
2024-09-10 17:00:20 +08:00
committed by GitHub
parent 178730266d
commit 2cf1187b32
724 changed files with 21180 additions and 21123 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
api/core/app/apps/advanced_chat/app_config_manager.py
View File

@@ -1,4 +1,3 @@
from core.app.app_config.base_app_config_manager import BaseAppConfigManager
from core.app.app_config.common.sensitive_word_avoidance.manager import SensitiveWordAvoidanceConfigManager
from core.app.app_config.entities import WorkflowUIBasedAppConfig
@@ -19,13 +18,13 @@ class AdvancedChatAppConfig(WorkflowUIBasedAppConfig):
"""
Advanced Chatbot App Config Entity.
"""
pass
class AdvancedChatAppConfigManager(BaseAppConfigManager):
@classmethod
def get_app_config(cls, app_model: App,
workflow: Workflow) -> AdvancedChatAppConfig:
def get_app_config(cls, app_model: App, workflow: Workflow) -> AdvancedChatAppConfig:
features_dict = workflow.features_dict
app_mode = AppMode.value_of(app_model.mode)
@@ -34,13 +33,9 @@ class AdvancedChatAppConfigManager(BaseAppConfigManager):
app_id=app_model.id,
app_mode=app_mode,
workflow_id=workflow.id,
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(
config=features_dict
),
variables=WorkflowVariablesConfigManager.convert(
workflow=workflow
),
additional_features=cls.convert_features(features_dict, app_mode)
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(config=features_dict),
variables=WorkflowVariablesConfigManager.convert(workflow=workflow),
additional_features=cls.convert_features(features_dict, app_mode),
)
return app_config
@@ -58,8 +53,7 @@ class AdvancedChatAppConfigManager(BaseAppConfigManager):
# file upload validation
config, current_related_config_keys = FileUploadConfigManager.validate_and_set_defaults(
config=config,
is_vision=False
config=config, is_vision=False
)
related_config_keys.extend(current_related_config_keys)
@@ -69,7 +63,8 @@ class AdvancedChatAppConfigManager(BaseAppConfigManager):
# suggested_questions_after_answer
config, current_related_config_keys = SuggestedQuestionsAfterAnswerConfigManager.validate_and_set_defaults(
config)
config
)
related_config_keys.extend(current_related_config_keys)
# speech_to_text
@@ -86,9 +81,7 @@ class AdvancedChatAppConfigManager(BaseAppConfigManager):
# moderation validation
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(
tenant_id=tenant_id,
config=config,
only_structure_validate=only_structure_validate
tenant_id=tenant_id, config=config, only_structure_validate=only_structure_validate
)
related_config_keys.extend(current_related_config_keys)
@@ -98,4 +91,3 @@ class AdvancedChatAppConfigManager(BaseAppConfigManager):
filtered_config = {key: config.get(key) for key in related_config_keys}
return filtered_config

166
api/core/app/apps/advanced_chat/app_generator.py
View File

@@ -34,7 +34,8 @@ logger = logging.getLogger(__name__)
class AdvancedChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
@@ -44,7 +45,8 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
@@ -53,14 +55,14 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
) -> dict: ...
def generate(
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: bool = True,
) -> dict[str, Any] | Generator[str, Any, None]:
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: bool = True,
) -> dict[str, Any] | Generator[str, Any, None]:
"""
Generate App response.
@@ -71,44 +73,37 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
:param invoke_from: invoke from source
:param stream: is stream
"""
if not args.get('query'):
raise ValueError('query is required')
if not args.get("query"):
raise ValueError("query is required")
query = args['query']
query = args["query"]
if not isinstance(query, str):
raise ValueError('query must be a string')
raise ValueError("query must be a string")
query = query.replace('\x00', '')
inputs = args['inputs']
query = query.replace("\x00", "")
inputs = args["inputs"]
extras = {
"auto_generate_conversation_name": args.get('auto_generate_name', False)
}
extras = {"auto_generate_conversation_name": args.get("auto_generate_name", False)}
# get conversation
conversation = None
conversation_id = args.get('conversation_id')
conversation_id = args.get("conversation_id")
if conversation_id:
conversation = self._get_conversation_by_user(app_model=app_model, conversation_id=conversation_id, user=user)
conversation = self._get_conversation_by_user(
app_model=app_model, conversation_id=conversation_id, user=user
)
# parse files
files = args['files'] if args.get('files') else []
files = args["files"] if args.get("files") else []
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(workflow.features_dict, is_vision=False)
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(files, file_extra_config, user)
else:
file_objs = []
# convert to app config
app_config = AdvancedChatAppConfigManager.get_app_config(
app_model=app_model,
workflow=workflow
)
app_config = AdvancedChatAppConfigManager.get_app_config(app_model=app_model, workflow=workflow)
# get tracing instance
user_id = user.id if isinstance(user, Account) else user.session_id
@@ -130,7 +125,7 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
stream=stream,
invoke_from=invoke_from,
extras=extras,
trace_manager=trace_manager
trace_manager=trace_manager,
)
contexts.tenant_id.set(application_generate_entity.app_config.tenant_id)
@@ -140,16 +135,12 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
invoke_from=invoke_from,
application_generate_entity=application_generate_entity,
conversation=conversation,
stream=stream
stream=stream,
)
def single_iteration_generate(self, app_model: App,
workflow: Workflow,
node_id: str,
user: Account,
args: dict,
stream: bool = True) \
-> dict[str, Any] | Generator[str, Any, None]:
def single_iteration_generate(
self, app_model: App, workflow: Workflow, node_id: str, user: Account, args: dict, stream: bool = True
) -> dict[str, Any] | Generator[str, Any, None]:
"""
Generate App response.
@@ -161,16 +152,13 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
:param stream: is stream
"""
if not node_id:
raise ValueError('node_id is required')
raise ValueError("node_id is required")
if args.get('inputs') is None:
raise ValueError('inputs is required')
if args.get("inputs") is None:
raise ValueError("inputs is required")
# convert to app config
app_config = AdvancedChatAppConfigManager.get_app_config(
app_model=app_model,
workflow=workflow
)
app_config = AdvancedChatAppConfigManager.get_app_config(app_model=app_model, workflow=workflow)
# init application generate entity
application_generate_entity = AdvancedChatAppGenerateEntity(
@@ -178,18 +166,15 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
app_config=app_config,
conversation_id=None,
inputs={},
query='',
query="",
files=[],
user_id=user.id,
stream=stream,
invoke_from=InvokeFrom.DEBUGGER,
extras={
"auto_generate_conversation_name": False
},
extras={"auto_generate_conversation_name": False},
single_iteration_run=AdvancedChatAppGenerateEntity.SingleIterationRunEntity(
node_id=node_id,
inputs=args['inputs']
)
node_id=node_id, inputs=args["inputs"]
),
)
contexts.tenant_id.set(application_generate_entity.app_config.tenant_id)
@@ -199,17 +184,19 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
invoke_from=InvokeFrom.DEBUGGER,
application_generate_entity=application_generate_entity,
conversation=None,
stream=stream
stream=stream,
)
def _generate(self, *,
workflow: Workflow,
user: Union[Account, EndUser],
invoke_from: InvokeFrom,
application_generate_entity: AdvancedChatAppGenerateEntity,
conversation: Optional[Conversation] = None,
stream: bool = True) \
-> dict[str, Any] | Generator[str, Any, None]:
def _generate(
self,
*,
workflow: Workflow,
user: Union[Account, EndUser],
invoke_from: InvokeFrom,
application_generate_entity: AdvancedChatAppGenerateEntity,
conversation: Optional[Conversation] = None,
stream: bool = True,
) -> dict[str, Any] | Generator[str, Any, None]:
"""
Generate App response.
@@ -225,10 +212,7 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
is_first_conversation = True
# init generate records
(
conversation,
message
) = self._init_generate_records(application_generate_entity, conversation)
(conversation, message) = self._init_generate_records(application_generate_entity, conversation)
if is_first_conversation:
# update conversation features
@@ -243,18 +227,21 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
invoke_from=application_generate_entity.invoke_from,
conversation_id=conversation.id,
app_mode=conversation.mode,
message_id=message.id
message_id=message.id,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(), # type: ignore
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'conversation_id': conversation.id,
'message_id': message.id,
'context': contextvars.copy_context(),
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(), # type: ignore
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"conversation_id": conversation.id,
"message_id": message.id,
"context": contextvars.copy_context(),
},
)
worker_thread.start()
@@ -269,17 +256,17 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
stream=stream,
)
return AdvancedChatAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return AdvancedChatAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
def _generate_worker(self, flask_app: Flask,
application_generate_entity: AdvancedChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation_id: str,
message_id: str,
context: contextvars.Context) -> None:
def _generate_worker(
self,
flask_app: Flask,
application_generate_entity: AdvancedChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation_id: str,
message_id: str,
context: contextvars.Context,
) -> None:
"""
Generate worker in a new thread.
:param flask_app: Flask app
@@ -302,7 +289,7 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
conversation=conversation,
message=message
message=message,
)
runner.run()
@@ -310,14 +297,13 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
pass
except InvokeAuthorizationError:
queue_manager.publish_error(
InvokeAuthorizationError('Incorrect API key provided'),
PublishFrom.APPLICATION_MANAGER
InvokeAuthorizationError("Incorrect API key provided"), PublishFrom.APPLICATION_MANAGER
)
except ValidationError as e:
logger.exception("Validation Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except (ValueError, InvokeError) as e:
if os.environ.get("DEBUG", "false").lower() == 'true':
if os.environ.get("DEBUG", "false").lower() == "true":
logger.exception("Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except Exception as e:

37
api/core/app/apps/advanced_chat/app_generator_tts_publisher.py
View File

@@ -25,10 +25,7 @@ def _invoiceTTS(text_content: str, model_instance, tenant_id: str, voice: str):
if not text_content or text_content.isspace():
return
return model_instance.invoke_tts(
content_text=text_content.strip(),
user="responding_tts",
tenant_id=tenant_id,
voice=voice
content_text=text_content.strip(), user="responding_tts", tenant_id=tenant_id, voice=voice
)
@@ -44,28 +41,26 @@ def _process_future(future_queue, audio_queue):
except Exception as e:
logging.getLogger(__name__).warning(e)
break
audio_queue.put(AudioTrunk("finish", b''))
audio_queue.put(AudioTrunk("finish", b""))
class AppGeneratorTTSPublisher:
def __init__(self, tenant_id: str, voice: str):
self.logger = logging.getLogger(__name__)
self.tenant_id = tenant_id
self.msg_text = ''
self.msg_text = ""
self._audio_queue = queue.Queue()
self._msg_queue = queue.Queue()
self.match = re.compile(r'[。.!?]')
self.match = re.compile(r"[。.!?]")
self.model_manager = ModelManager()
self.model_instance = self.model_manager.get_default_model_instance(
tenant_id=self.tenant_id,
model_type=ModelType.TTS
tenant_id=self.tenant_id, model_type=ModelType.TTS
)
self.voices = self.model_instance.get_tts_voices()
values = [voice.get('value') for voice in self.voices]
values = [voice.get("value") for voice in self.voices]
self.voice = voice
if not voice or voice not in values:
self.voice = self.voices[0].get('value')
self.voice = self.voices[0].get("value")
self.MAX_SENTENCE = 2
self._last_audio_event = None
self._runtime_thread = threading.Thread(target=self._runtime).start()
@@ -85,8 +80,9 @@ class AppGeneratorTTSPublisher:
message = self._msg_queue.get()
if message is None:
if self.msg_text and len(self.msg_text.strip()) > 0:
futures_result = self.executor.submit(_invoiceTTS, self.msg_text,
self.model_instance, self.tenant_id, self.voice)
futures_result = self.executor.submit(
_invoiceTTS, self.msg_text, self.model_instance, self.tenant_id, self.voice
)
future_queue.put(futures_result)
break
elif isinstance(message.event, QueueAgentMessageEvent | QueueLLMChunkEvent):
@@ -94,21 +90,20 @@ class AppGeneratorTTSPublisher:
elif isinstance(message.event, QueueTextChunkEvent):
self.msg_text += message.event.text
elif isinstance(message.event, QueueNodeSucceededEvent):
self.msg_text += message.event.outputs.get('output', '')
self.msg_text += message.event.outputs.get("output", "")
self.last_message = message
sentence_arr, text_tmp = self._extract_sentence(self.msg_text)
if len(sentence_arr) >= min(self.MAX_SENTENCE, 7):
self.MAX_SENTENCE += 1
text_content = ''.join(sentence_arr)
futures_result = self.executor.submit(_invoiceTTS, text_content,
self.model_instance,
self.tenant_id,
self.voice)
text_content = "".join(sentence_arr)
futures_result = self.executor.submit(
_invoiceTTS, text_content, self.model_instance, self.tenant_id, self.voice
)
future_queue.put(futures_result)
if text_tmp:
self.msg_text = text_tmp
else:
self.msg_text = ''
self.msg_text = ""
except Exception as e:
self.logger.warning(e)

84
api/core/app/apps/advanced_chat/app_runner.py
View File

@@ -38,11 +38,11 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
"""
def __init__(
self,
application_generate_entity: AdvancedChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message
self,
application_generate_entity: AdvancedChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
) -> None:
"""
:param application_generate_entity: application generate entity
@@ -66,11 +66,11 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
app_record = db.session.query(App).filter(App.id == app_config.app_id).first()
if not app_record:
raise ValueError('App not found')
raise ValueError("App not found")
workflow = self.get_workflow(app_model=app_record, workflow_id=app_config.workflow_id)
if not workflow:
raise ValueError('Workflow not initialized')
raise ValueError("Workflow not initialized")
user_id = None
if self.application_generate_entity.invoke_from in [InvokeFrom.WEB_APP, InvokeFrom.SERVICE_API]:
@@ -81,7 +81,7 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
user_id = self.application_generate_entity.user_id
workflow_callbacks: list[WorkflowCallback] = []
if bool(os.environ.get("DEBUG", 'False').lower() == 'true'):
if bool(os.environ.get("DEBUG", "False").lower() == "true"):
workflow_callbacks.append(WorkflowLoggingCallback())
if self.application_generate_entity.single_iteration_run:
@@ -89,7 +89,7 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
graph, variable_pool = self._get_graph_and_variable_pool_of_single_iteration(
workflow=workflow,
node_id=self.application_generate_entity.single_iteration_run.node_id,
user_inputs=self.application_generate_entity.single_iteration_run.inputs
user_inputs=self.application_generate_entity.single_iteration_run.inputs,
)
else:
inputs = self.application_generate_entity.inputs
@@ -98,26 +98,27 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
# moderation
if self.handle_input_moderation(
app_record=app_record,
app_generate_entity=self.application_generate_entity,
inputs=inputs,
query=query,
message_id=self.message.id
app_record=app_record,
app_generate_entity=self.application_generate_entity,
inputs=inputs,
query=query,
message_id=self.message.id,
):
return
# annotation reply
if self.handle_annotation_reply(
app_record=app_record,
message=self.message,
query=query,
app_generate_entity=self.application_generate_entity
app_record=app_record,
message=self.message,
query=query,
app_generate_entity=self.application_generate_entity,
):
return
# Init conversation variables
stmt = select(ConversationVariable).where(
ConversationVariable.app_id == self.conversation.app_id, ConversationVariable.conversation_id == self.conversation.id
ConversationVariable.app_id == self.conversation.app_id,
ConversationVariable.conversation_id == self.conversation.id,
)
with Session(db.engine) as session:
conversation_variables = session.scalars(stmt).all()
@@ -190,12 +191,12 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
self._handle_event(workflow_entry, event)
def handle_input_moderation(
self,
app_record: App,
app_generate_entity: AdvancedChatAppGenerateEntity,
inputs: Mapping[str, Any],
query: str,
message_id: str
self,
app_record: App,
app_generate_entity: AdvancedChatAppGenerateEntity,
inputs: Mapping[str, Any],
query: str,
message_id: str,
) -> bool:
"""
Handle input moderation
@@ -217,18 +218,14 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
message_id=message_id,
)
except ModerationException as e:
self._complete_with_stream_output(
text=str(e),
stopped_by=QueueStopEvent.StopBy.INPUT_MODERATION
)
self._complete_with_stream_output(text=str(e), stopped_by=QueueStopEvent.StopBy.INPUT_MODERATION)
return True
return False
def handle_annotation_reply(self, app_record: App,
message: Message,
query: str,
app_generate_entity: AdvancedChatAppGenerateEntity) -> bool:
def handle_annotation_reply(
self, app_record: App, message: Message, query: str, app_generate_entity: AdvancedChatAppGenerateEntity
) -> bool:
"""
Handle annotation reply
:param app_record: app record
@@ -246,32 +243,21 @@ class AdvancedChatAppRunner(WorkflowBasedAppRunner):
)
if annotation_reply:
self._publish_event(
QueueAnnotationReplyEvent(message_annotation_id=annotation_reply.id)
)
self._publish_event(QueueAnnotationReplyEvent(message_annotation_id=annotation_reply.id))
self._complete_with_stream_output(
text=annotation_reply.content,
stopped_by=QueueStopEvent.StopBy.ANNOTATION_REPLY
text=annotation_reply.content, stopped_by=QueueStopEvent.StopBy.ANNOTATION_REPLY
)
return True
return False
def _complete_with_stream_output(self,
text: str,
stopped_by: QueueStopEvent.StopBy) -> None:
def _complete_with_stream_output(self, text: str, stopped_by: QueueStopEvent.StopBy) -> None:
"""
Direct output
:param text: text
:return:
"""
self._publish_event(
QueueTextChunkEvent(
text=text
)
)
self._publish_event(QueueTextChunkEvent(text=text))
self._publish_event(
QueueStopEvent(stopped_by=stopped_by)
)
self._publish_event(QueueStopEvent(stopped_by=stopped_by))

54
api/core/app/apps/advanced_chat/generate_response_converter.py
View File

@@ -28,15 +28,15 @@ class AdvancedChatAppGenerateResponseConverter(AppGenerateResponseConverter):
"""
blocking_response = cast(ChatbotAppBlockingResponse, blocking_response)
response = {
'event': 'message',
'task_id': blocking_response.task_id,
'id': blocking_response.data.id,
'message_id': blocking_response.data.message_id,
'conversation_id': blocking_response.data.conversation_id,
'mode': blocking_response.data.mode,
'answer': blocking_response.data.answer,
'metadata': blocking_response.data.metadata,
'created_at': blocking_response.data.created_at
"event": "message",
"task_id": blocking_response.task_id,
"id": blocking_response.data.id,
"message_id": blocking_response.data.message_id,
"conversation_id": blocking_response.data.conversation_id,
"mode": blocking_response.data.mode,
"answer": blocking_response.data.answer,
"metadata": blocking_response.data.metadata,
"created_at": blocking_response.data.created_at,
}
return response
@@ -50,13 +50,15 @@ class AdvancedChatAppGenerateResponseConverter(AppGenerateResponseConverter):
"""
response = cls.convert_blocking_full_response(blocking_response)
metadata = response.get('metadata', {})
response['metadata'] = cls._get_simple_metadata(metadata)
metadata = response.get("metadata", {})
response["metadata"] = cls._get_simple_metadata(metadata)
return response
@classmethod
def convert_stream_full_response(cls, stream_response: Generator[AppStreamResponse, None, None]) -> Generator[str, Any, None]:
def convert_stream_full_response(
cls, stream_response: Generator[AppStreamResponse, None, None]
) -> Generator[str, Any, None]:
"""
Convert stream full response.
:param stream_response: stream response
@@ -67,14 +69,14 @@ class AdvancedChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, ErrorStreamResponse):
@@ -85,7 +87,9 @@ class AdvancedChatAppGenerateResponseConverter(AppGenerateResponseConverter):
yield json.dumps(response_chunk)
@classmethod
def convert_stream_simple_response(cls, stream_response: Generator[AppStreamResponse, None, None]) -> Generator[str, Any, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[AppStreamResponse, None, None]
) -> Generator[str, Any, None]:
"""
Convert stream simple response.
:param stream_response: stream response
@@ -96,20 +100,20 @@ class AdvancedChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, MessageEndStreamResponse):
sub_stream_response_dict = sub_stream_response.to_dict()
metadata = sub_stream_response_dict.get('metadata', {})
sub_stream_response_dict['metadata'] = cls._get_simple_metadata(metadata)
metadata = sub_stream_response_dict.get("metadata", {})
sub_stream_response_dict["metadata"] = cls._get_simple_metadata(metadata)
response_chunk.update(sub_stream_response_dict)
if isinstance(sub_stream_response, ErrorStreamResponse):
data = cls._error_to_stream_response(sub_stream_response.err)

167
api/core/app/apps/advanced_chat/generate_task_pipeline.py
View File

@@ -65,6 +65,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
"""
AdvancedChatAppGenerateTaskPipeline is a class that generate stream output and state management for Application.
"""
_task_state: WorkflowTaskState
_application_generate_entity: AdvancedChatAppGenerateEntity
_workflow: Workflow
@@ -72,14 +73,14 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
_workflow_system_variables: dict[SystemVariableKey, Any]
def __init__(
self,
application_generate_entity: AdvancedChatAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
user: Union[Account, EndUser],
stream: bool,
self,
application_generate_entity: AdvancedChatAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
user: Union[Account, EndUser],
stream: bool,
) -> None:
"""
Initialize AdvancedChatAppGenerateTaskPipeline.
@@ -123,13 +124,10 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
# start generate conversation name thread
self._conversation_name_generate_thread = self._generate_conversation_name(
self._conversation,
self._application_generate_entity.query
self._conversation, self._application_generate_entity.query
)
generator = self._wrapper_process_stream_response(
trace_manager=self._application_generate_entity.trace_manager
)
generator = self._wrapper_process_stream_response(trace_manager=self._application_generate_entity.trace_manager)
if self._stream:
return self._to_stream_response(generator)
@@ -147,7 +145,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
elif isinstance(stream_response, MessageEndStreamResponse):
extras = {}
if stream_response.metadata:
extras['metadata'] = stream_response.metadata
extras["metadata"] = stream_response.metadata
return ChatbotAppBlockingResponse(
task_id=stream_response.task_id,
@@ -158,15 +156,17 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
message_id=self._message.id,
answer=self._task_state.answer,
created_at=int(self._message.created_at.timestamp()),
**extras
)
**extras,
),
)
else:
continue
raise Exception('Queue listening stopped unexpectedly.')
raise Exception("Queue listening stopped unexpectedly.")
def _to_stream_response(self, generator: Generator[StreamResponse, None, None]) -> Generator[ChatbotAppStreamResponse, Any, None]:
def _to_stream_response(
self, generator: Generator[StreamResponse, None, None]
) -> Generator[ChatbotAppStreamResponse, Any, None]:
"""
To stream response.
:return:
@@ -176,7 +176,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
conversation_id=self._conversation.id,
message_id=self._message.id,
created_at=int(self._message.created_at.timestamp()),
stream_response=stream_response
stream_response=stream_response,
)
def _listenAudioMsg(self, publisher, task_id: str):
@@ -187,17 +187,20 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
return MessageAudioStreamResponse(audio=audio_msg.audio, task_id=task_id)
return None
def _wrapper_process_stream_response(self, trace_manager: Optional[TraceQueueManager] = None) -> \
Generator[StreamResponse, None, None]:
def _wrapper_process_stream_response(
self, trace_manager: Optional[TraceQueueManager] = None
) -> Generator[StreamResponse, None, None]:
tts_publisher = None
task_id = self._application_generate_entity.task_id
tenant_id = self._application_generate_entity.app_config.tenant_id
features_dict = self._workflow.features_dict
if features_dict.get('text_to_speech') and features_dict['text_to_speech'].get('enabled') and features_dict[
'text_to_speech'].get('autoPlay') == 'enabled':
tts_publisher = AppGeneratorTTSPublisher(tenant_id, features_dict['text_to_speech'].get('voice'))
if (
features_dict.get("text_to_speech")
and features_dict["text_to_speech"].get("enabled")
and features_dict["text_to_speech"].get("autoPlay") == "enabled"
):
tts_publisher = AppGeneratorTTSPublisher(tenant_id, features_dict["text_to_speech"].get("voice"))
for response in self._process_stream_response(tts_publisher=tts_publisher, trace_manager=trace_manager):
while True:
@@ -228,12 +231,12 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
except Exception as e:
logger.error(e)
break
yield MessageAudioEndStreamResponse(audio='', task_id=task_id)
yield MessageAudioEndStreamResponse(audio="", task_id=task_id)
def _process_stream_response(
self,
tts_publisher: Optional[AppGeneratorTTSPublisher] = None,
trace_manager: Optional[TraceQueueManager] = None
self,
tts_publisher: Optional[AppGeneratorTTSPublisher] = None,
trace_manager: Optional[TraceQueueManager] = None,
) -> Generator[StreamResponse, None, None]:
"""
Process stream response.
@@ -267,22 +270,18 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
db.session.close()
yield self._workflow_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
elif isinstance(event, QueueNodeStartedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
workflow_node_execution = self._handle_node_execution_start(
workflow_run=workflow_run,
event=event
)
workflow_node_execution = self._handle_node_execution_start(workflow_run=workflow_run, event=event)
response = self._workflow_node_start_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
@@ -293,7 +292,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
response = self._workflow_node_finish_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
@@ -304,62 +303,52 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
response = self._workflow_node_finish_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
yield response
elif isinstance(event, QueueParallelBranchRunStartedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_parallel_branch_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueParallelBranchRunSucceededEvent | QueueParallelBranchRunFailedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_parallel_branch_finished_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationStartEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationNextEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_next_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationCompletedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_completed_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueWorkflowSucceededEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
if not graph_runtime_state:
raise Exception('Graph runtime state not initialized.')
raise Exception("Graph runtime state not initialized.")
workflow_run = self._handle_workflow_run_success(
workflow_run=workflow_run,
@@ -372,20 +361,16 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
)
yield self._workflow_finish_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
self._queue_manager.publish(
QueueAdvancedChatMessageEndEvent(),
PublishFrom.TASK_PIPELINE
)
self._queue_manager.publish(QueueAdvancedChatMessageEndEvent(), PublishFrom.TASK_PIPELINE)
elif isinstance(event, QueueWorkflowFailedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
if not graph_runtime_state:
raise Exception('Graph runtime state not initialized.')
raise Exception("Graph runtime state not initialized.")
workflow_run = self._handle_workflow_run_failed(
workflow_run=workflow_run,
@@ -399,11 +384,10 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
)
yield self._workflow_finish_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
err_event = QueueErrorEvent(error=ValueError(f'Run failed: {workflow_run.error}'))
err_event = QueueErrorEvent(error=ValueError(f"Run failed: {workflow_run.error}"))
yield self._error_to_stream_response(self._handle_error(err_event, self._message))
break
elif isinstance(event, QueueStopEvent):
@@ -420,8 +404,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
)
yield self._workflow_finish_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
# Save message
@@ -434,8 +417,9 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
self._refetch_message()
self._message.message_metadata = json.dumps(jsonable_encoder(self._task_state.metadata)) \
if self._task_state.metadata else None
self._message.message_metadata = (
json.dumps(jsonable_encoder(self._task_state.metadata)) if self._task_state.metadata else None
)
db.session.commit()
db.session.refresh(self._message)
@@ -445,8 +429,9 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
self._refetch_message()
self._message.message_metadata = json.dumps(jsonable_encoder(self._task_state.metadata)) \
if self._task_state.metadata else None
self._message.message_metadata = (
json.dumps(jsonable_encoder(self._task_state.metadata)) if self._task_state.metadata else None
)
db.session.commit()
db.session.refresh(self._message)
@@ -472,7 +457,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
yield self._message_replace_to_stream_response(answer=event.text)
elif isinstance(event, QueueAdvancedChatMessageEndEvent):
if not graph_runtime_state:
raise Exception('Graph runtime state not initialized.')
raise Exception("Graph runtime state not initialized.")
output_moderation_answer = self._handle_output_moderation_when_task_finished(self._task_state.answer)
if output_moderation_answer:
@@ -502,8 +487,9 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
self._message.answer = self._task_state.answer
self._message.provider_response_latency = time.perf_counter() - self._start_at
self._message.message_metadata = json.dumps(jsonable_encoder(self._task_state.metadata)) \
if self._task_state.metadata else None
self._message.message_metadata = (
json.dumps(jsonable_encoder(self._task_state.metadata)) if self._task_state.metadata else None
)
if graph_runtime_state and graph_runtime_state.llm_usage:
usage = graph_runtime_state.llm_usage
@@ -523,7 +509,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
application_generate_entity=self._application_generate_entity,
conversation=self._conversation,
is_first_message=self._application_generate_entity.conversation_id is None,
extras=self._application_generate_entity.extras
extras=self._application_generate_entity.extras,
)
def _message_end_to_stream_response(self) -> MessageEndStreamResponse:
@@ -533,15 +519,13 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
"""
extras = {}
if self._task_state.metadata:
extras['metadata'] = self._task_state.metadata.copy()
extras["metadata"] = self._task_state.metadata.copy()
if 'annotation_reply' in extras['metadata']:
del extras['metadata']['annotation_reply']
if "annotation_reply" in extras["metadata"]:
del extras["metadata"]["annotation_reply"]
return MessageEndStreamResponse(
task_id=self._application_generate_entity.task_id,
id=self._message.id,
**extras
task_id=self._application_generate_entity.task_id, id=self._message.id, **extras
)
def _handle_output_moderation_chunk(self, text: str) -> bool:
@@ -555,14 +539,11 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
# stop subscribe new token when output moderation should direct output
self._task_state.answer = self._output_moderation_handler.get_final_output()
self._queue_manager.publish(
QueueTextChunkEvent(
text=self._task_state.answer
), PublishFrom.TASK_PIPELINE
QueueTextChunkEvent(text=self._task_state.answer), PublishFrom.TASK_PIPELINE
)
self._queue_manager.publish(
QueueStopEvent(stopped_by=QueueStopEvent.StopBy.OUTPUT_MODERATION),
PublishFrom.TASK_PIPELINE
QueueStopEvent(stopped_by=QueueStopEvent.StopBy.OUTPUT_MODERATION), PublishFrom.TASK_PIPELINE
)
return True
else:

59
api/core/app/apps/agent_chat/app_config_manager.py
View File

@@ -28,15 +28,19 @@ class AgentChatAppConfig(EasyUIBasedAppConfig):
"""
Agent Chatbot App Config Entity.
"""
agent: Optional[AgentEntity] = None
class AgentChatAppConfigManager(BaseAppConfigManager):
@classmethod
def get_app_config(cls, app_model: App,
app_model_config: AppModelConfig,
conversation: Optional[Conversation] = None,
override_config_dict: Optional[dict] = None) -> AgentChatAppConfig:
def get_app_config(
cls,
app_model: App,
app_model_config: AppModelConfig,
conversation: Optional[Conversation] = None,
override_config_dict: Optional[dict] = None,
) -> AgentChatAppConfig:
"""
Convert app model config to agent chat app config
:param app_model: app model
@@ -66,22 +70,12 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
app_model_config_from=config_from,
app_model_config_id=app_model_config.id,
app_model_config_dict=config_dict,
model=ModelConfigManager.convert(
config=config_dict
),
prompt_template=PromptTemplateConfigManager.convert(
config=config_dict
),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(
config=config_dict
),
dataset=DatasetConfigManager.convert(
config=config_dict
),
agent=AgentConfigManager.convert(
config=config_dict
),
additional_features=cls.convert_features(config_dict, app_mode)
model=ModelConfigManager.convert(config=config_dict),
prompt_template=PromptTemplateConfigManager.convert(config=config_dict),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(config=config_dict),
dataset=DatasetConfigManager.convert(config=config_dict),
agent=AgentConfigManager.convert(config=config_dict),
additional_features=cls.convert_features(config_dict, app_mode),
)
app_config.variables, app_config.external_data_variables = BasicVariablesConfigManager.convert(
@@ -128,7 +122,8 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
# suggested_questions_after_answer
config, current_related_config_keys = SuggestedQuestionsAfterAnswerConfigManager.validate_and_set_defaults(
config)
config
)
related_config_keys.extend(current_related_config_keys)
# speech_to_text
@@ -145,13 +140,15 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
# dataset configs
# dataset_query_variable
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(tenant_id, app_mode,
config)
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(
tenant_id, app_mode, config
)
related_config_keys.extend(current_related_config_keys)
# moderation validation
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(tenant_id,
config)
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(
tenant_id, config
)
related_config_keys.extend(current_related_config_keys)
related_config_keys = list(set(related_config_keys))
@@ -170,10 +167,7 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
:param config: app model config args
"""
if not config.get("agent_mode"):
config["agent_mode"] = {
"enabled": False,
"tools": []
}
config["agent_mode"] = {"enabled": False, "tools": []}
if not isinstance(config["agent_mode"], dict):
raise ValueError("agent_mode must be of object type")
@@ -187,8 +181,9 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
if not config["agent_mode"].get("strategy"):
config["agent_mode"]["strategy"] = PlanningStrategy.ROUTER.value
if config["agent_mode"]["strategy"] not in [member.value for member in
list(PlanningStrategy.__members__.values())]:
if config["agent_mode"]["strategy"] not in [
member.value for member in list(PlanningStrategy.__members__.values())
]:
raise ValueError("strategy in agent_mode must be in the specified strategy list")
if not config["agent_mode"].get("tools"):
@@ -210,7 +205,7 @@ class AgentChatAppConfigManager(BaseAppConfigManager):
raise ValueError("enabled in agent_mode.tools must be of boolean type")
if key == "dataset":
if 'id' not in tool_item:
if "id" not in tool_item:
raise ValueError("id is required in dataset")
try:

102
api/core/app/apps/agent_chat/app_generator.py
View File

@@ -30,7 +30,8 @@ logger = logging.getLogger(__name__)
class AgentChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
@@ -39,19 +40,17 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: Literal[False] = False,
) -> dict: ...
def generate(self, app_model: App,
user: Union[Account, EndUser],
args: Any,
invoke_from: InvokeFrom,
stream: bool = True) \
-> Union[dict, Generator[dict, None, None]]:
def generate(
self, app_model: App, user: Union[Account, EndUser], args: Any, invoke_from: InvokeFrom, stream: bool = True
) -> Union[dict, Generator[dict, None, None]]:
"""
Generate App response.
@@ -62,60 +61,48 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
:param stream: is stream
"""
if not stream:
raise ValueError('Agent Chat App does not support blocking mode')
raise ValueError("Agent Chat App does not support blocking mode")
if not args.get('query'):
raise ValueError('query is required')
if not args.get("query"):
raise ValueError("query is required")
query = args['query']
query = args["query"]
if not isinstance(query, str):
raise ValueError('query must be a string')
raise ValueError("query must be a string")
query = query.replace('\x00', '')
inputs = args['inputs']
query = query.replace("\x00", "")
inputs = args["inputs"]
extras = {
"auto_generate_conversation_name": args.get('auto_generate_name', True)
}
extras = {"auto_generate_conversation_name": args.get("auto_generate_name", True)}
# get conversation
conversation = None
if args.get('conversation_id'):
conversation = self._get_conversation_by_user(app_model, args.get('conversation_id'), user)
if args.get("conversation_id"):
conversation = self._get_conversation_by_user(app_model, args.get("conversation_id"), user)
# get app model config
app_model_config = self._get_app_model_config(
app_model=app_model,
conversation=conversation
)
app_model_config = self._get_app_model_config(app_model=app_model, conversation=conversation)
# validate override model config
override_model_config_dict = None
if args.get('model_config'):
if args.get("model_config"):
if invoke_from != InvokeFrom.DEBUGGER:
raise ValueError('Only in App debug mode can override model config')
raise ValueError("Only in App debug mode can override model config")
# validate config
override_model_config_dict = AgentChatAppConfigManager.config_validate(
tenant_id=app_model.tenant_id,
config=args.get('model_config')
tenant_id=app_model.tenant_id, config=args.get("model_config")
)
# always enable retriever resource in debugger mode
override_model_config_dict["retriever_resource"] = {
"enabled": True
}
override_model_config_dict["retriever_resource"] = {"enabled": True}
# parse files
files = args['files'] if args.get('files') else []
files = args["files"] if args.get("files") else []
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(files, file_extra_config, user)
else:
file_objs = []
@@ -124,7 +111,7 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
app_model=app_model,
app_model_config=app_model_config,
conversation=conversation,
override_config_dict=override_model_config_dict
override_config_dict=override_model_config_dict,
)
# get tracing instance
@@ -145,14 +132,11 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
invoke_from=invoke_from,
extras=extras,
call_depth=0,
trace_manager=trace_manager
trace_manager=trace_manager,
)
# init generate records
(
conversation,
message
) = self._init_generate_records(application_generate_entity, conversation)
(conversation, message) = self._init_generate_records(application_generate_entity, conversation)
# init queue manager
queue_manager = MessageBasedAppQueueManager(
@@ -161,17 +145,20 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
invoke_from=application_generate_entity.invoke_from,
conversation_id=conversation.id,
app_mode=conversation.mode,
message_id=message.id
message_id=message.id,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(),
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'conversation_id': conversation.id,
'message_id': message.id,
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(),
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"conversation_id": conversation.id,
"message_id": message.id,
},
)
worker_thread.start()
@@ -185,13 +172,11 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
stream=stream,
)
return AgentChatAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return AgentChatAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
def _generate_worker(
self, flask_app: Flask,
self,
flask_app: Flask,
application_generate_entity: AgentChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation_id: str,
@@ -224,14 +209,13 @@ class AgentChatAppGenerator(MessageBasedAppGenerator):
pass
except InvokeAuthorizationError:
queue_manager.publish_error(
InvokeAuthorizationError('Incorrect API key provided'),
PublishFrom.APPLICATION_MANAGER
InvokeAuthorizationError("Incorrect API key provided"), PublishFrom.APPLICATION_MANAGER
)
except ValidationError as e:
logger.exception("Validation Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except (ValueError, InvokeError) as e:
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == 'true':
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == "true":
logger.exception("Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except Exception as e:

91
api/core/app/apps/agent_chat/app_runner.py
View File

@@ -30,7 +30,8 @@ class AgentChatAppRunner(AppRunner):
"""
def run(
self, application_generate_entity: AgentChatAppGenerateEntity,
self,
application_generate_entity: AgentChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
@@ -65,7 +66,7 @@ class AgentChatAppRunner(AppRunner):
prompt_template_entity=app_config.prompt_template,
inputs=inputs,
files=files,
query=query
query=query,
)
memory = None
@@ -73,13 +74,10 @@ class AgentChatAppRunner(AppRunner):
# get memory of conversation (read-only)
model_instance = ModelInstance(
provider_model_bundle=application_generate_entity.model_conf.provider_model_bundle,
model=application_generate_entity.model_conf.model
model=application_generate_entity.model_conf.model,
)
memory = TokenBufferMemory(
conversation=conversation,
model_instance=model_instance
)
memory = TokenBufferMemory(conversation=conversation, model_instance=model_instance)
# organize all inputs and template to prompt messages
# Include: prompt template, inputs, query(optional), files(optional)
@@ -91,7 +89,7 @@ class AgentChatAppRunner(AppRunner):
inputs=inputs,
files=files,
query=query,
memory=memory
memory=memory,
)
# moderation
@@ -103,7 +101,7 @@ class AgentChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
inputs=inputs,
query=query,
message_id=message.id
message_id=message.id,
)
except ModerationException as e:
self.direct_output(
@@ -111,7 +109,7 @@ class AgentChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text=str(e),
stream=application_generate_entity.stream
stream=application_generate_entity.stream,
)
return
@@ -122,13 +120,13 @@ class AgentChatAppRunner(AppRunner):
message=message,
query=query,
user_id=application_generate_entity.user_id,
invoke_from=application_generate_entity.invoke_from
invoke_from=application_generate_entity.invoke_from,
)
if annotation_reply:
queue_manager.publish(
QueueAnnotationReplyEvent(message_annotation_id=annotation_reply.id),
PublishFrom.APPLICATION_MANAGER
PublishFrom.APPLICATION_MANAGER,
)
self.direct_output(
@@ -136,7 +134,7 @@ class AgentChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text=annotation_reply.content,
stream=application_generate_entity.stream
stream=application_generate_entity.stream,
)
return
@@ -148,7 +146,7 @@ class AgentChatAppRunner(AppRunner):
app_id=app_record.id,
external_data_tools=external_data_tools,
inputs=inputs,
query=query
query=query,
)
# reorganize all inputs and template to prompt messages
@@ -161,14 +159,14 @@ class AgentChatAppRunner(AppRunner):
inputs=inputs,
files=files,
query=query,
memory=memory
memory=memory,
)
# check hosting moderation
hosting_moderation_result = self.check_hosting_moderation(
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
prompt_messages=prompt_messages
prompt_messages=prompt_messages,
)
if hosting_moderation_result:
@@ -177,9 +175,9 @@ class AgentChatAppRunner(AppRunner):
agent_entity = app_config.agent
# load tool variables
tool_conversation_variables = self._load_tool_variables(conversation_id=conversation.id,
user_id=application_generate_entity.user_id,
tenant_id=app_config.tenant_id)
tool_conversation_variables = self._load_tool_variables(
conversation_id=conversation.id, user_id=application_generate_entity.user_id, tenant_id=app_config.tenant_id
)
# convert db variables to tool variables
tool_variables = self._convert_db_variables_to_tool_variables(tool_conversation_variables)
@@ -187,7 +185,7 @@ class AgentChatAppRunner(AppRunner):
# init model instance
model_instance = ModelInstance(
provider_model_bundle=application_generate_entity.model_conf.provider_model_bundle,
model=application_generate_entity.model_conf.model
model=application_generate_entity.model_conf.model,
)
prompt_message, _ = self.organize_prompt_messages(
app_record=app_record,
@@ -238,7 +236,7 @@ class AgentChatAppRunner(AppRunner):
prompt_messages=prompt_message,
variables_pool=tool_variables,
db_variables=tool_conversation_variables,
model_instance=model_instance
model_instance=model_instance,
)
invoke_result = runner.run(
@@ -252,17 +250,21 @@ class AgentChatAppRunner(AppRunner):
invoke_result=invoke_result,
queue_manager=queue_manager,
stream=application_generate_entity.stream,
agent=True
agent=True,
)
def _load_tool_variables(self, conversation_id: str, user_id: str, tenant_id: str) -> ToolConversationVariables:
"""
load tool variables from database
"""
tool_variables: ToolConversationVariables = db.session.query(ToolConversationVariables).filter(
ToolConversationVariables.conversation_id == conversation_id,
ToolConversationVariables.tenant_id == tenant_id
).first()
tool_variables: ToolConversationVariables = (
db.session.query(ToolConversationVariables)
.filter(
ToolConversationVariables.conversation_id == conversation_id,
ToolConversationVariables.tenant_id == tenant_id,
)
.first()
)
if tool_variables:
# save tool variables to session, so that we can update it later
@@ -273,34 +275,40 @@ class AgentChatAppRunner(AppRunner):
conversation_id=conversation_id,
user_id=user_id,
tenant_id=tenant_id,
variables_str='[]',
variables_str="[]",
)
db.session.add(tool_variables)
db.session.commit()
return tool_variables
def _convert_db_variables_to_tool_variables(self, db_variables: ToolConversationVariables) -> ToolRuntimeVariablePool:
def _convert_db_variables_to_tool_variables(
self, db_variables: ToolConversationVariables
) -> ToolRuntimeVariablePool:
"""
convert db variables to tool variables
"""
return ToolRuntimeVariablePool(**{
'conversation_id': db_variables.conversation_id,
'user_id': db_variables.user_id,
'tenant_id': db_variables.tenant_id,
'pool': db_variables.variables
})
return ToolRuntimeVariablePool(
**{
"conversation_id": db_variables.conversation_id,
"user_id": db_variables.user_id,
"tenant_id": db_variables.tenant_id,
"pool": db_variables.variables,
}
)
def _get_usage_of_all_agent_thoughts(self, model_config: ModelConfigWithCredentialsEntity,
message: Message) -> LLMUsage:
def _get_usage_of_all_agent_thoughts(
self, model_config: ModelConfigWithCredentialsEntity, message: Message
) -> LLMUsage:
"""
Get usage of all agent thoughts
:param model_config: model config
:param message: message
:return:
"""
agent_thoughts = (db.session.query(MessageAgentThought)
.filter(MessageAgentThought.message_id == message.id).all())
agent_thoughts = (
db.session.query(MessageAgentThought).filter(MessageAgentThought.message_id == message.id).all()
)
all_message_tokens = 0
all_answer_tokens = 0
@@ -312,8 +320,5 @@ class AgentChatAppRunner(AppRunner):
model_type_instance = cast(LargeLanguageModel, model_type_instance)
return model_type_instance._calc_response_usage(
model_config.model,
model_config.credentials,
all_message_tokens,
all_answer_tokens
model_config.model, model_config.credentials, all_message_tokens, all_answer_tokens
)

56
api/core/app/apps/agent_chat/generate_response_converter.py
View File

@@ -23,15 +23,15 @@ class AgentChatAppGenerateResponseConverter(AppGenerateResponseConverter):
:return:
"""
response = {
'event': 'message',
'task_id': blocking_response.task_id,
'id': blocking_response.data.id,
'message_id': blocking_response.data.message_id,
'conversation_id': blocking_response.data.conversation_id,
'mode': blocking_response.data.mode,
'answer': blocking_response.data.answer,
'metadata': blocking_response.data.metadata,
'created_at': blocking_response.data.created_at
"event": "message",
"task_id": blocking_response.task_id,
"id": blocking_response.data.id,
"message_id": blocking_response.data.message_id,
"conversation_id": blocking_response.data.conversation_id,
"mode": blocking_response.data.mode,
"answer": blocking_response.data.answer,
"metadata": blocking_response.data.metadata,
"created_at": blocking_response.data.created_at,
}
return response
@@ -45,14 +45,15 @@ class AgentChatAppGenerateResponseConverter(AppGenerateResponseConverter):
"""
response = cls.convert_blocking_full_response(blocking_response)
metadata = response.get('metadata', {})
response['metadata'] = cls._get_simple_metadata(metadata)
metadata = response.get("metadata", {})
response["metadata"] = cls._get_simple_metadata(metadata)
return response
@classmethod
def convert_stream_full_response(cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_full_response(
cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream full response.
:param stream_response: stream response
@@ -63,14 +64,14 @@ class AgentChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, ErrorStreamResponse):
@@ -81,8 +82,9 @@ class AgentChatAppGenerateResponseConverter(AppGenerateResponseConverter):
yield json.dumps(response_chunk)
@classmethod
def convert_stream_simple_response(cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream simple response.
:param stream_response: stream response
@@ -93,20 +95,20 @@ class AgentChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, MessageEndStreamResponse):
sub_stream_response_dict = sub_stream_response.to_dict()
metadata = sub_stream_response_dict.get('metadata', {})
sub_stream_response_dict['metadata'] = cls._get_simple_metadata(metadata)
metadata = sub_stream_response_dict.get("metadata", {})
sub_stream_response_dict["metadata"] = cls._get_simple_metadata(metadata)
response_chunk.update(sub_stream_response_dict)
if isinstance(sub_stream_response, ErrorStreamResponse):
data = cls._error_to_stream_response(sub_stream_response.err)

85
api/core/app/apps/base_app_generate_response_converter.py
View File

@@ -13,32 +13,33 @@ class AppGenerateResponseConverter(ABC):
_blocking_response_type: type[AppBlockingResponse]
@classmethod
def convert(cls, response: Union[
AppBlockingResponse,
Generator[AppStreamResponse, Any, None]
], invoke_from: InvokeFrom) -> dict[str, Any] | Generator[str, Any, None]:
def convert(
cls, response: Union[AppBlockingResponse, Generator[AppStreamResponse, Any, None]], invoke_from: InvokeFrom
) -> dict[str, Any] | Generator[str, Any, None]:
if invoke_from in [InvokeFrom.DEBUGGER, InvokeFrom.SERVICE_API]:
if isinstance(response, AppBlockingResponse):
return cls.convert_blocking_full_response(response)
else:
def _generate_full_response() -> Generator[str, Any, None]:
for chunk in cls.convert_stream_full_response(response):
if chunk == 'ping':
yield f'event: {chunk}\n\n'
if chunk == "ping":
yield f"event: {chunk}\n\n"
else:
yield f'data: {chunk}\n\n'
yield f"data: {chunk}\n\n"
return _generate_full_response()
else:
if isinstance(response, AppBlockingResponse):
return cls.convert_blocking_simple_response(response)
else:
def _generate_simple_response() -> Generator[str, Any, None]:
for chunk in cls.convert_stream_simple_response(response):
if chunk == 'ping':
yield f'event: {chunk}\n\n'
if chunk == "ping":
yield f"event: {chunk}\n\n"
else:
yield f'data: {chunk}\n\n'
yield f"data: {chunk}\n\n"
return _generate_simple_response()
@@ -54,14 +55,16 @@ class AppGenerateResponseConverter(ABC):
@classmethod
@abstractmethod
def convert_stream_full_response(cls, stream_response: Generator[AppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_full_response(
cls, stream_response: Generator[AppStreamResponse, None, None]
) -> Generator[str, None, None]:
raise NotImplementedError
@classmethod
@abstractmethod
def convert_stream_simple_response(cls, stream_response: Generator[AppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[AppStreamResponse, None, None]
) -> Generator[str, None, None]:
raise NotImplementedError
@classmethod
@@ -72,24 +75,26 @@ class AppGenerateResponseConverter(ABC):
:return:
"""
# show_retrieve_source
if 'retriever_resources' in metadata:
metadata['retriever_resources'] = []
for resource in metadata['retriever_resources']:
metadata['retriever_resources'].append({
'segment_id': resource['segment_id'],
'position': resource['position'],
'document_name': resource['document_name'],
'score': resource['score'],
'content': resource['content'],
})
if "retriever_resources" in metadata:
metadata["retriever_resources"] = []
for resource in metadata["retriever_resources"]:
metadata["retriever_resources"].append(
{
"segment_id": resource["segment_id"],
"position": resource["position"],
"document_name": resource["document_name"],
"score": resource["score"],
"content": resource["content"],
}
)
# show annotation reply
if 'annotation_reply' in metadata:
del metadata['annotation_reply']
if "annotation_reply" in metadata:
del metadata["annotation_reply"]
# show usage
if 'usage' in metadata:
del metadata['usage']
if "usage" in metadata:
del metadata["usage"]
return metadata
@@ -101,16 +106,16 @@ class AppGenerateResponseConverter(ABC):
:return:
"""
error_responses = {
ValueError: {'code': 'invalid_param', 'status': 400},
ProviderTokenNotInitError: {'code': 'provider_not_initialize', 'status': 400},
ValueError: {"code": "invalid_param", "status": 400},
ProviderTokenNotInitError: {"code": "provider_not_initialize", "status": 400},
QuotaExceededError: {
'code': 'provider_quota_exceeded',
'message': "Your quota for Dify Hosted Model Provider has been exhausted. "
"Please go to Settings -> Model Provider to complete your own provider credentials.",
'status': 400
"code": "provider_quota_exceeded",
"message": "Your quota for Dify Hosted Model Provider has been exhausted. "
"Please go to Settings -> Model Provider to complete your own provider credentials.",
"status": 400,
},
ModelCurrentlyNotSupportError: {'code': 'model_currently_not_support', 'status': 400},
InvokeError: {'code': 'completion_request_error', 'status': 400}
ModelCurrentlyNotSupportError: {"code": "model_currently_not_support", "status": 400},
InvokeError: {"code": "completion_request_error", "status": 400},
}
# Determine the response based on the type of exception
@@ -120,13 +125,13 @@ class AppGenerateResponseConverter(ABC):
data = v
if data:
data.setdefault('message', getattr(e, 'description', str(e)))
data.setdefault("message", getattr(e, "description", str(e)))
else:
logging.error(e)
data = {
'code': 'internal_server_error',
'message': 'Internal Server Error, please contact support.',
'status': 500
"code": "internal_server_error",
"message": "Internal Server Error, please contact support.",
"status": 500,
}
return data

12
api/core/app/apps/base_app_generator.py
View File

@@ -16,10 +16,10 @@ class BaseAppGenerator:
def _validate_input(self, *, inputs: Mapping[str, Any], var: VariableEntity):
user_input_value = inputs.get(var.variable)
if var.required and not user_input_value:
raise ValueError(f'{var.variable} is required in input form')
raise ValueError(f"{var.variable} is required in input form")
if not var.required and not user_input_value:
# TODO: should we return None here if the default value is None?
return var.default or ''
return var.default or ""
if (
var.type
in (
@@ -34,7 +34,7 @@ class BaseAppGenerator:
if var.type == VariableEntityType.NUMBER and isinstance(user_input_value, str):
# may raise ValueError if user_input_value is not a valid number
try:
if '.' in user_input_value:
if "." in user_input_value:
return float(user_input_value)
else:
return int(user_input_value)
@@ -43,14 +43,14 @@ class BaseAppGenerator:
if var.type == VariableEntityType.SELECT:
options = var.options or []
if user_input_value not in options:
raise ValueError(f'{var.variable} in input form must be one of the following: {options}')
raise ValueError(f"{var.variable} in input form must be one of the following: {options}")
elif var.type in (VariableEntityType.TEXT_INPUT, VariableEntityType.PARAGRAPH):
if var.max_length and user_input_value and len(user_input_value) > var.max_length:
raise ValueError(f'{var.variable} in input form must be less than {var.max_length} characters')
raise ValueError(f"{var.variable} in input form must be less than {var.max_length} characters")
return user_input_value
def _sanitize_value(self, value: Any) -> Any:
if isinstance(value, str):
return value.replace('\x00', '')
return value.replace("\x00", "")
return value

30
api/core/app/apps/base_app_queue_manager.py
View File

@@ -24,9 +24,7 @@ class PublishFrom(Enum):
class AppQueueManager:
def __init__(self, task_id: str,
user_id: str,
invoke_from: InvokeFrom) -> None:
def __init__(self, task_id: str, user_id: str, invoke_from: InvokeFrom) -> None:
if not user_id:
raise ValueError("user is required")
@@ -34,9 +32,10 @@ class AppQueueManager:
self._user_id = user_id
self._invoke_from = invoke_from
user_prefix = 'account' if self._invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER] else 'end-user'
redis_client.setex(AppQueueManager._generate_task_belong_cache_key(self._task_id), 1800,
f"{user_prefix}-{self._user_id}")
user_prefix = "account" if self._invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER] else "end-user"
redis_client.setex(
AppQueueManager._generate_task_belong_cache_key(self._task_id), 1800, f"{user_prefix}-{self._user_id}"
)
q = queue.Queue()
@@ -66,8 +65,7 @@ class AppQueueManager:
# publish two messages to make sure the client can receive the stop signal
# and stop listening after the stop signal processed
self.publish(
QueueStopEvent(stopped_by=QueueStopEvent.StopBy.USER_MANUAL),
PublishFrom.TASK_PIPELINE
QueueStopEvent(stopped_by=QueueStopEvent.StopBy.USER_MANUAL), PublishFrom.TASK_PIPELINE
)
if elapsed_time // 10 > last_ping_time:
@@ -88,9 +86,7 @@ class AppQueueManager:
:param pub_from: publish from
:return:
"""
self.publish(QueueErrorEvent(
error=e
), pub_from)
self.publish(QueueErrorEvent(error=e), pub_from)
def publish(self, event: AppQueueEvent, pub_from: PublishFrom) -> None:
"""
@@ -122,8 +118,8 @@ class AppQueueManager:
if result is None:
return
user_prefix = 'account' if invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER] else 'end-user'
if result.decode('utf-8') != f"{user_prefix}-{user_id}":
user_prefix = "account" if invoke_from in [InvokeFrom.EXPLORE, InvokeFrom.DEBUGGER] else "end-user"
if result.decode("utf-8") != f"{user_prefix}-{user_id}":
return
stopped_cache_key = cls._generate_stopped_cache_key(task_id)
@@ -168,9 +164,11 @@ class AppQueueManager:
for item in data:
self._check_for_sqlalchemy_models(item)
else:
if isinstance(data, DeclarativeMeta) or hasattr(data, '_sa_instance_state'):
raise TypeError("Critical Error: Passing SQLAlchemy Model instances "
"that cause thread safety issues is not allowed.")
if isinstance(data, DeclarativeMeta) or hasattr(data, "_sa_instance_state"):
raise TypeError(
"Critical Error: Passing SQLAlchemy Model instances "
"that cause thread safety issues is not allowed."
)
class GenerateTaskStoppedException(Exception):

272
api/core/app/apps/base_app_runner.py
View File

@@ -31,12 +31,15 @@ if TYPE_CHECKING:
class AppRunner:
def get_pre_calculate_rest_tokens(self, app_record: App,
model_config: ModelConfigWithCredentialsEntity,
prompt_template_entity: PromptTemplateEntity,
inputs: dict[str, str],
files: list["FileVar"],
query: Optional[str] = None) -> int:
def get_pre_calculate_rest_tokens(
self,
app_record: App,
model_config: ModelConfigWithCredentialsEntity,
prompt_template_entity: PromptTemplateEntity,
inputs: dict[str, str],
files: list["FileVar"],
query: Optional[str] = None,
) -> int:
"""
Get pre calculate rest tokens
:param app_record: app record
@@ -49,18 +52,20 @@ class AppRunner:
"""
# Invoke model
model_instance = ModelInstance(
provider_model_bundle=model_config.provider_model_bundle,
model=model_config.model
provider_model_bundle=model_config.provider_model_bundle, model=model_config.model
)
model_context_tokens = model_config.model_schema.model_properties.get(ModelPropertyKey.CONTEXT_SIZE)
max_tokens = 0
for parameter_rule in model_config.model_schema.parameter_rules:
if (parameter_rule.name == 'max_tokens'
or (parameter_rule.use_template and parameter_rule.use_template == 'max_tokens')):
max_tokens = (model_config.parameters.get(parameter_rule.name)
or model_config.parameters.get(parameter_rule.use_template)) or 0
if parameter_rule.name == "max_tokens" or (
parameter_rule.use_template and parameter_rule.use_template == "max_tokens"
):
max_tokens = (
model_config.parameters.get(parameter_rule.name)
or model_config.parameters.get(parameter_rule.use_template)
) or 0
if model_context_tokens is None:
return -1
@@ -75,36 +80,39 @@ class AppRunner:
prompt_template_entity=prompt_template_entity,
inputs=inputs,
files=files,
query=query
query=query,
)
prompt_tokens = model_instance.get_llm_num_tokens(
prompt_messages
)
prompt_tokens = model_instance.get_llm_num_tokens(prompt_messages)
rest_tokens = model_context_tokens - max_tokens - prompt_tokens
if rest_tokens < 0:
raise InvokeBadRequestError("Query or prefix prompt is too long, you can reduce the prefix prompt, "
"or shrink the max token, or switch to a llm with a larger token limit size.")
raise InvokeBadRequestError(
"Query or prefix prompt is too long, you can reduce the prefix prompt, "
"or shrink the max token, or switch to a llm with a larger token limit size."
)
return rest_tokens
def recalc_llm_max_tokens(self, model_config: ModelConfigWithCredentialsEntity,
prompt_messages: list[PromptMessage]):
def recalc_llm_max_tokens(
self, model_config: ModelConfigWithCredentialsEntity, prompt_messages: list[PromptMessage]
):
# recalc max_tokens if sum(prompt_token + max_tokens) over model token limit
model_instance = ModelInstance(
provider_model_bundle=model_config.provider_model_bundle,
model=model_config.model
provider_model_bundle=model_config.provider_model_bundle, model=model_config.model
)
model_context_tokens = model_config.model_schema.model_properties.get(ModelPropertyKey.CONTEXT_SIZE)
max_tokens = 0
for parameter_rule in model_config.model_schema.parameter_rules:
if (parameter_rule.name == 'max_tokens'
or (parameter_rule.use_template and parameter_rule.use_template == 'max_tokens')):
max_tokens = (model_config.parameters.get(parameter_rule.name)
or model_config.parameters.get(parameter_rule.use_template)) or 0
if parameter_rule.name == "max_tokens" or (
parameter_rule.use_template and parameter_rule.use_template == "max_tokens"
):
max_tokens = (
model_config.parameters.get(parameter_rule.name)
or model_config.parameters.get(parameter_rule.use_template)
) or 0
if model_context_tokens is None:
return -1
@@ -112,27 +120,28 @@ class AppRunner:
if max_tokens is None:
max_tokens = 0
prompt_tokens = model_instance.get_llm_num_tokens(
prompt_messages
)
prompt_tokens = model_instance.get_llm_num_tokens(prompt_messages)
if prompt_tokens + max_tokens > model_context_tokens:
max_tokens = max(model_context_tokens - prompt_tokens, 16)
for parameter_rule in model_config.model_schema.parameter_rules:
if (parameter_rule.name == 'max_tokens'
or (parameter_rule.use_template and parameter_rule.use_template == 'max_tokens')):
if parameter_rule.name == "max_tokens" or (
parameter_rule.use_template and parameter_rule.use_template == "max_tokens"
):
model_config.parameters[parameter_rule.name] = max_tokens
def organize_prompt_messages(self, app_record: App,
model_config: ModelConfigWithCredentialsEntity,
prompt_template_entity: PromptTemplateEntity,
inputs: dict[str, str],
files: list["FileVar"],
query: Optional[str] = None,
context: Optional[str] = None,
memory: Optional[TokenBufferMemory] = None) \
-> tuple[list[PromptMessage], Optional[list[str]]]:
def organize_prompt_messages(
self,
app_record: App,
model_config: ModelConfigWithCredentialsEntity,
prompt_template_entity: PromptTemplateEntity,
inputs: dict[str, str],
files: list["FileVar"],
query: Optional[str] = None,
context: Optional[str] = None,
memory: Optional[TokenBufferMemory] = None,
) -> tuple[list[PromptMessage], Optional[list[str]]]:
"""
Organize prompt messages
:param context:
@@ -152,60 +161,54 @@ class AppRunner:
app_mode=AppMode.value_of(app_record.mode),
prompt_template_entity=prompt_template_entity,
inputs=inputs,
query=query if query else '',
query=query if query else "",
files=files,
context=context,
memory=memory,
model_config=model_config
model_config=model_config,
)
else:
memory_config = MemoryConfig(
window=MemoryConfig.WindowConfig(
enabled=False
)
)
memory_config = MemoryConfig(window=MemoryConfig.WindowConfig(enabled=False))
model_mode = ModelMode.value_of(model_config.mode)
if model_mode == ModelMode.COMPLETION:
advanced_completion_prompt_template = prompt_template_entity.advanced_completion_prompt_template
prompt_template = CompletionModelPromptTemplate(
text=advanced_completion_prompt_template.prompt
)
prompt_template = CompletionModelPromptTemplate(text=advanced_completion_prompt_template.prompt)
if advanced_completion_prompt_template.role_prefix:
memory_config.role_prefix = MemoryConfig.RolePrefix(
user=advanced_completion_prompt_template.role_prefix.user,
assistant=advanced_completion_prompt_template.role_prefix.assistant
assistant=advanced_completion_prompt_template.role_prefix.assistant,
)
else:
prompt_template = []
for message in prompt_template_entity.advanced_chat_prompt_template.messages:
prompt_template.append(ChatModelMessage(
text=message.text,
role=message.role
))
prompt_template.append(ChatModelMessage(text=message.text, role=message.role))
prompt_transform = AdvancedPromptTransform()
prompt_messages = prompt_transform.get_prompt(
prompt_template=prompt_template,
inputs=inputs,
query=query if query else '',
query=query if query else "",
files=files,
context=context,
memory_config=memory_config,
memory=memory,
model_config=model_config
model_config=model_config,
)
stop = model_config.stop
return prompt_messages, stop
def direct_output(self, queue_manager: AppQueueManager,
app_generate_entity: EasyUIBasedAppGenerateEntity,
prompt_messages: list,
text: str,
stream: bool,
usage: Optional[LLMUsage] = None) -> None:
def direct_output(
self,
queue_manager: AppQueueManager,
app_generate_entity: EasyUIBasedAppGenerateEntity,
prompt_messages: list,
text: str,
stream: bool,
usage: Optional[LLMUsage] = None,
) -> None:
"""
Direct output
:param queue_manager: application queue manager
@@ -222,17 +225,10 @@ class AppRunner:
chunk = LLMResultChunk(
model=app_generate_entity.model_conf.model,
prompt_messages=prompt_messages,
delta=LLMResultChunkDelta(
index=index,
message=AssistantPromptMessage(content=token)
)
delta=LLMResultChunkDelta(index=index, message=AssistantPromptMessage(content=token)),
)
queue_manager.publish(
QueueLLMChunkEvent(
chunk=chunk
), PublishFrom.APPLICATION_MANAGER
)
queue_manager.publish(QueueLLMChunkEvent(chunk=chunk), PublishFrom.APPLICATION_MANAGER)
index += 1
time.sleep(0.01)
@@ -242,15 +238,19 @@ class AppRunner:
model=app_generate_entity.model_conf.model,
prompt_messages=prompt_messages,
message=AssistantPromptMessage(content=text),
usage=usage if usage else LLMUsage.empty_usage()
usage=usage if usage else LLMUsage.empty_usage(),
),
), PublishFrom.APPLICATION_MANAGER
),
PublishFrom.APPLICATION_MANAGER,
)
def _handle_invoke_result(self, invoke_result: Union[LLMResult, Generator],
queue_manager: AppQueueManager,
stream: bool,
agent: bool = False) -> None:
def _handle_invoke_result(
self,
invoke_result: Union[LLMResult, Generator],
queue_manager: AppQueueManager,
stream: bool,
agent: bool = False,
) -> None:
"""
Handle invoke result
:param invoke_result: invoke result
@@ -260,21 +260,13 @@ class AppRunner:
:return:
"""
if not stream:
self._handle_invoke_result_direct(
invoke_result=invoke_result,
queue_manager=queue_manager,
agent=agent
)
self._handle_invoke_result_direct(invoke_result=invoke_result, queue_manager=queue_manager, agent=agent)
else:
self._handle_invoke_result_stream(
invoke_result=invoke_result,
queue_manager=queue_manager,
agent=agent
)
self._handle_invoke_result_stream(invoke_result=invoke_result, queue_manager=queue_manager, agent=agent)
def _handle_invoke_result_direct(self, invoke_result: LLMResult,
queue_manager: AppQueueManager,
agent: bool) -> None:
def _handle_invoke_result_direct(
self, invoke_result: LLMResult, queue_manager: AppQueueManager, agent: bool
) -> None:
"""
Handle invoke result direct
:param invoke_result: invoke result
@@ -285,12 +277,13 @@ class AppRunner:
queue_manager.publish(
QueueMessageEndEvent(
llm_result=invoke_result,
), PublishFrom.APPLICATION_MANAGER
),
PublishFrom.APPLICATION_MANAGER,
)
def _handle_invoke_result_stream(self, invoke_result: Generator,
queue_manager: AppQueueManager,
agent: bool) -> None:
def _handle_invoke_result_stream(
self, invoke_result: Generator, queue_manager: AppQueueManager, agent: bool
) -> None:
"""
Handle invoke result
:param invoke_result: invoke result
@@ -300,21 +293,13 @@ class AppRunner:
"""
model = None
prompt_messages = []
text = ''
text = ""
usage = None
for result in invoke_result:
if not agent:
queue_manager.publish(
QueueLLMChunkEvent(
chunk=result
), PublishFrom.APPLICATION_MANAGER
)
queue_manager.publish(QueueLLMChunkEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
else:
queue_manager.publish(
QueueAgentMessageEvent(
chunk=result
), PublishFrom.APPLICATION_MANAGER
)
queue_manager.publish(QueueAgentMessageEvent(chunk=result), PublishFrom.APPLICATION_MANAGER)
text += result.delta.message.content
@@ -331,25 +316,24 @@ class AppRunner:
usage = LLMUsage.empty_usage()
llm_result = LLMResult(
model=model,
prompt_messages=prompt_messages,
message=AssistantPromptMessage(content=text),
usage=usage
model=model, prompt_messages=prompt_messages, message=AssistantPromptMessage(content=text), usage=usage
)
queue_manager.publish(
QueueMessageEndEvent(
llm_result=llm_result,
), PublishFrom.APPLICATION_MANAGER
),
PublishFrom.APPLICATION_MANAGER,
)
def moderation_for_inputs(
self, app_id: str,
tenant_id: str,
app_generate_entity: AppGenerateEntity,
inputs: Mapping[str, Any],
query: str,
message_id: str,
self,
app_id: str,
tenant_id: str,
app_generate_entity: AppGenerateEntity,
inputs: Mapping[str, Any],
query: str,
message_id: str,
) -> tuple[bool, dict, str]:
"""
Process sensitive_word_avoidance.
@@ -367,14 +351,17 @@ class AppRunner:
tenant_id=tenant_id,
app_config=app_generate_entity.app_config,
inputs=inputs,
query=query if query else '',
query=query if query else "",
message_id=message_id,
trace_manager=app_generate_entity.trace_manager
trace_manager=app_generate_entity.trace_manager,
)
def check_hosting_moderation(self, application_generate_entity: EasyUIBasedAppGenerateEntity,
queue_manager: AppQueueManager,
prompt_messages: list[PromptMessage]) -> bool:
def check_hosting_moderation(
self,
application_generate_entity: EasyUIBasedAppGenerateEntity,
queue_manager: AppQueueManager,
prompt_messages: list[PromptMessage],
) -> bool:
"""
Check hosting moderation
:param application_generate_entity: application generate entity
@@ -384,8 +371,7 @@ class AppRunner:
"""
hosting_moderation_feature = HostingModerationFeature()
moderation_result = hosting_moderation_feature.check(
application_generate_entity=application_generate_entity,
prompt_messages=prompt_messages
application_generate_entity=application_generate_entity, prompt_messages=prompt_messages
)
if moderation_result:
@@ -393,18 +379,20 @@ class AppRunner:
queue_manager=queue_manager,
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text="I apologize for any confusion, " \
"but I'm an AI assistant to be helpful, harmless, and honest.",
stream=application_generate_entity.stream
text="I apologize for any confusion, " "but I'm an AI assistant to be helpful, harmless, and honest.",
stream=application_generate_entity.stream,
)
return moderation_result
def fill_in_inputs_from_external_data_tools(self, tenant_id: str,
app_id: str,
external_data_tools: list[ExternalDataVariableEntity],
inputs: dict,
query: str) -> dict:
def fill_in_inputs_from_external_data_tools(
self,
tenant_id: str,
app_id: str,
external_data_tools: list[ExternalDataVariableEntity],
inputs: dict,
query: str,
) -> dict:
"""
Fill in variable inputs from external data tools if exists.
@@ -417,18 +405,12 @@ class AppRunner:
"""
external_data_fetch_feature = ExternalDataFetch()
return external_data_fetch_feature.fetch(
tenant_id=tenant_id,
app_id=app_id,
external_data_tools=external_data_tools,
inputs=inputs,
query=query
tenant_id=tenant_id, app_id=app_id, external_data_tools=external_data_tools, inputs=inputs, query=query
)
def query_app_annotations_to_reply(self, app_record: App,
message: Message,
query: str,
user_id: str,
invoke_from: InvokeFrom) -> Optional[MessageAnnotation]:
def query_app_annotations_to_reply(
self, app_record: App, message: Message, query: str, user_id: str, invoke_from: InvokeFrom
) -> Optional[MessageAnnotation]:
"""
Query app annotations to reply
:param app_record: app record
@@ -440,9 +422,5 @@ class AppRunner:
"""
annotation_reply_feature = AnnotationReplyFeature()
return annotation_reply_feature.query(
app_record=app_record,
message=message,
query=query,
user_id=user_id,
invoke_from=invoke_from
app_record=app_record, message=message, query=query, user_id=user_id, invoke_from=invoke_from
)

45
api/core/app/apps/chat/app_config_manager.py
View File

@@ -22,15 +22,19 @@ class ChatAppConfig(EasyUIBasedAppConfig):
"""
Chatbot App Config Entity.
"""
pass
class ChatAppConfigManager(BaseAppConfigManager):
@classmethod
def get_app_config(cls, app_model: App,
app_model_config: AppModelConfig,
conversation: Optional[Conversation] = None,
override_config_dict: Optional[dict] = None) -> ChatAppConfig:
def get_app_config(
cls,
app_model: App,
app_model_config: AppModelConfig,
conversation: Optional[Conversation] = None,
override_config_dict: Optional[dict] = None,
) -> ChatAppConfig:
"""
Convert app model config to chat app config
:param app_model: app model
@@ -51,7 +55,7 @@ class ChatAppConfigManager(BaseAppConfigManager):
config_dict = app_model_config_dict.copy()
else:
if not override_config_dict:
raise Exception('override_config_dict is required when config_from is ARGS')
raise Exception("override_config_dict is required when config_from is ARGS")
config_dict = override_config_dict
@@ -63,19 +67,11 @@ class ChatAppConfigManager(BaseAppConfigManager):
app_model_config_from=config_from,
app_model_config_id=app_model_config.id,
app_model_config_dict=config_dict,
model=ModelConfigManager.convert(
config=config_dict
),
prompt_template=PromptTemplateConfigManager.convert(
config=config_dict
),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(
config=config_dict
),
dataset=DatasetConfigManager.convert(
config=config_dict
),
additional_features=cls.convert_features(config_dict, app_mode)
model=ModelConfigManager.convert(config=config_dict),
prompt_template=PromptTemplateConfigManager.convert(config=config_dict),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(config=config_dict),
dataset=DatasetConfigManager.convert(config=config_dict),
additional_features=cls.convert_features(config_dict, app_mode),
)
app_config.variables, app_config.external_data_variables = BasicVariablesConfigManager.convert(
@@ -113,8 +109,9 @@ class ChatAppConfigManager(BaseAppConfigManager):
related_config_keys.extend(current_related_config_keys)
# dataset_query_variable
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(tenant_id, app_mode,
config)
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(
tenant_id, app_mode, config
)
related_config_keys.extend(current_related_config_keys)
# opening_statement
@@ -123,7 +120,8 @@ class ChatAppConfigManager(BaseAppConfigManager):
# suggested_questions_after_answer
config, current_related_config_keys = SuggestedQuestionsAfterAnswerConfigManager.validate_and_set_defaults(
config)
config
)
related_config_keys.extend(current_related_config_keys)
# speech_to_text
@@ -139,8 +137,9 @@ class ChatAppConfigManager(BaseAppConfigManager):
related_config_keys.extend(current_related_config_keys)
# moderation validation
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(tenant_id,
config)
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(
tenant_id, config
)
related_config_keys.extend(current_related_config_keys)
related_config_keys = list(set(related_config_keys))

106
api/core/app/apps/chat/app_generator.py
View File

@@ -30,7 +30,8 @@ logger = logging.getLogger(__name__)
class ChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: Any,
invoke_from: InvokeFrom,
@@ -39,7 +40,8 @@ class ChatAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: Any,
invoke_from: InvokeFrom,
@@ -47,7 +49,8 @@ class ChatAppGenerator(MessageBasedAppGenerator):
) -> dict: ...
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: Any,
invoke_from: InvokeFrom,
@@ -62,58 +65,46 @@ class ChatAppGenerator(MessageBasedAppGenerator):
:param invoke_from: invoke from source
:param stream: is stream
"""
if not args.get('query'):
raise ValueError('query is required')
if not args.get("query"):
raise ValueError("query is required")
query = args['query']
query = args["query"]
if not isinstance(query, str):
raise ValueError('query must be a string')
raise ValueError("query must be a string")
query = query.replace('\x00', '')
inputs = args['inputs']
query = query.replace("\x00", "")
inputs = args["inputs"]
extras = {
"auto_generate_conversation_name": args.get('auto_generate_name', True)
}
extras = {"auto_generate_conversation_name": args.get("auto_generate_name", True)}
# get conversation
conversation = None
if args.get('conversation_id'):
conversation = self._get_conversation_by_user(app_model, args.get('conversation_id'), user)
if args.get("conversation_id"):
conversation = self._get_conversation_by_user(app_model, args.get("conversation_id"), user)
# get app model config
app_model_config = self._get_app_model_config(
app_model=app_model,
conversation=conversation
)
app_model_config = self._get_app_model_config(app_model=app_model, conversation=conversation)
# validate override model config
override_model_config_dict = None
if args.get('model_config'):
if args.get("model_config"):
if invoke_from != InvokeFrom.DEBUGGER:
raise ValueError('Only in App debug mode can override model config')
raise ValueError("Only in App debug mode can override model config")
# validate config
override_model_config_dict = ChatAppConfigManager.config_validate(
tenant_id=app_model.tenant_id,
config=args.get('model_config')
tenant_id=app_model.tenant_id, config=args.get("model_config")
)
# always enable retriever resource in debugger mode
override_model_config_dict["retriever_resource"] = {
"enabled": True
}
override_model_config_dict["retriever_resource"] = {"enabled": True}
# parse files
files = args['files'] if args.get('files') else []
files = args["files"] if args.get("files") else []
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(files, file_extra_config, user)
else:
file_objs = []
@@ -122,7 +113,7 @@ class ChatAppGenerator(MessageBasedAppGenerator):
app_model=app_model,
app_model_config=app_model_config,
conversation=conversation,
override_config_dict=override_model_config_dict
override_config_dict=override_model_config_dict,
)
# get tracing instance
@@ -141,14 +132,11 @@ class ChatAppGenerator(MessageBasedAppGenerator):
stream=stream,
invoke_from=invoke_from,
extras=extras,
trace_manager=trace_manager
trace_manager=trace_manager,
)
# init generate records
(
conversation,
message
) = self._init_generate_records(application_generate_entity, conversation)
(conversation, message) = self._init_generate_records(application_generate_entity, conversation)
# init queue manager
queue_manager = MessageBasedAppQueueManager(
@@ -157,17 +145,20 @@ class ChatAppGenerator(MessageBasedAppGenerator):
invoke_from=application_generate_entity.invoke_from,
conversation_id=conversation.id,
app_mode=conversation.mode,
message_id=message.id
message_id=message.id,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(),
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'conversation_id': conversation.id,
'message_id': message.id,
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(),
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"conversation_id": conversation.id,
"message_id": message.id,
},
)
worker_thread.start()
@@ -181,16 +172,16 @@ class ChatAppGenerator(MessageBasedAppGenerator):
stream=stream,
)
return ChatAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return ChatAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
def _generate_worker(self, flask_app: Flask,
application_generate_entity: ChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation_id: str,
message_id: str) -> None:
def _generate_worker(
self,
flask_app: Flask,
application_generate_entity: ChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation_id: str,
message_id: str,
) -> None:
"""
Generate worker in a new thread.
:param flask_app: Flask app
@@ -212,20 +203,19 @@ class ChatAppGenerator(MessageBasedAppGenerator):
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
conversation=conversation,
message=message
message=message,
)
except GenerateTaskStoppedException:
pass
except InvokeAuthorizationError:
queue_manager.publish_error(
InvokeAuthorizationError('Incorrect API key provided'),
PublishFrom.APPLICATION_MANAGER
InvokeAuthorizationError("Incorrect API key provided"), PublishFrom.APPLICATION_MANAGER
)
except ValidationError as e:
logger.exception("Validation Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except (ValueError, InvokeError) as e:
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == 'true':
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == "true":
logger.exception("Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except Exception as e:

51
api/core/app/apps/chat/app_runner.py
View File

@@ -24,10 +24,13 @@ class ChatAppRunner(AppRunner):
Chat Application Runner
"""
def run(self, application_generate_entity: ChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message) -> None:
def run(
self,
application_generate_entity: ChatAppGenerateEntity,
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
) -> None:
"""
Run application
:param application_generate_entity: application generate entity
@@ -58,7 +61,7 @@ class ChatAppRunner(AppRunner):
prompt_template_entity=app_config.prompt_template,
inputs=inputs,
files=files,
query=query
query=query,
)
memory = None
@@ -66,13 +69,10 @@ class ChatAppRunner(AppRunner):
# get memory of conversation (read-only)
model_instance = ModelInstance(
provider_model_bundle=application_generate_entity.model_conf.provider_model_bundle,
model=application_generate_entity.model_conf.model
model=application_generate_entity.model_conf.model,
)
memory = TokenBufferMemory(
conversation=conversation,
model_instance=model_instance
)
memory = TokenBufferMemory(conversation=conversation, model_instance=model_instance)
# organize all inputs and template to prompt messages
# Include: prompt template, inputs, query(optional), files(optional)
@@ -84,7 +84,7 @@ class ChatAppRunner(AppRunner):
inputs=inputs,
files=files,
query=query,
memory=memory
memory=memory,
)
# moderation
@@ -96,7 +96,7 @@ class ChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
inputs=inputs,
query=query,
message_id=message.id
message_id=message.id,
)
except ModerationException as e:
self.direct_output(
@@ -104,7 +104,7 @@ class ChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text=str(e),
stream=application_generate_entity.stream
stream=application_generate_entity.stream,
)
return
@@ -115,13 +115,13 @@ class ChatAppRunner(AppRunner):
message=message,
query=query,
user_id=application_generate_entity.user_id,
invoke_from=application_generate_entity.invoke_from
invoke_from=application_generate_entity.invoke_from,
)
if annotation_reply:
queue_manager.publish(
QueueAnnotationReplyEvent(message_annotation_id=annotation_reply.id),
PublishFrom.APPLICATION_MANAGER
PublishFrom.APPLICATION_MANAGER,
)
self.direct_output(
@@ -129,7 +129,7 @@ class ChatAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text=annotation_reply.content,
stream=application_generate_entity.stream
stream=application_generate_entity.stream,
)
return
@@ -141,7 +141,7 @@ class ChatAppRunner(AppRunner):
app_id=app_record.id,
external_data_tools=external_data_tools,
inputs=inputs,
query=query
query=query,
)
# get context from datasets
@@ -152,7 +152,7 @@ class ChatAppRunner(AppRunner):
app_record.id,
message.id,
application_generate_entity.user_id,
application_generate_entity.invoke_from
application_generate_entity.invoke_from,
)
dataset_retrieval = DatasetRetrieval(application_generate_entity)
@@ -181,29 +181,26 @@ class ChatAppRunner(AppRunner):
files=files,
query=query,
context=context,
memory=memory
memory=memory,
)
# check hosting moderation
hosting_moderation_result = self.check_hosting_moderation(
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
prompt_messages=prompt_messages
prompt_messages=prompt_messages,
)
if hosting_moderation_result:
return
# Re-calculate the max tokens if sum(prompt_token + max_tokens) over model token limit
self.recalc_llm_max_tokens(
model_config=application_generate_entity.model_conf,
prompt_messages=prompt_messages
)
self.recalc_llm_max_tokens(model_config=application_generate_entity.model_conf, prompt_messages=prompt_messages)
# Invoke model
model_instance = ModelInstance(
provider_model_bundle=application_generate_entity.model_conf.provider_model_bundle,
model=application_generate_entity.model_conf.model
model=application_generate_entity.model_conf.model,
)
db.session.close()
@@ -218,7 +215,5 @@ class ChatAppRunner(AppRunner):
# handle invoke result
self._handle_invoke_result(
invoke_result=invoke_result,
queue_manager=queue_manager,
stream=application_generate_entity.stream
invoke_result=invoke_result, queue_manager=queue_manager, stream=application_generate_entity.stream
)

56
api/core/app/apps/chat/generate_response_converter.py
View File

@@ -23,15 +23,15 @@ class ChatAppGenerateResponseConverter(AppGenerateResponseConverter):
:return:
"""
response = {
'event': 'message',
'task_id': blocking_response.task_id,
'id': blocking_response.data.id,
'message_id': blocking_response.data.message_id,
'conversation_id': blocking_response.data.conversation_id,
'mode': blocking_response.data.mode,
'answer': blocking_response.data.answer,
'metadata': blocking_response.data.metadata,
'created_at': blocking_response.data.created_at
"event": "message",
"task_id": blocking_response.task_id,
"id": blocking_response.data.id,
"message_id": blocking_response.data.message_id,
"conversation_id": blocking_response.data.conversation_id,
"mode": blocking_response.data.mode,
"answer": blocking_response.data.answer,
"metadata": blocking_response.data.metadata,
"created_at": blocking_response.data.created_at,
}
return response
@@ -45,14 +45,15 @@ class ChatAppGenerateResponseConverter(AppGenerateResponseConverter):
"""
response = cls.convert_blocking_full_response(blocking_response)
metadata = response.get('metadata', {})
response['metadata'] = cls._get_simple_metadata(metadata)
metadata = response.get("metadata", {})
response["metadata"] = cls._get_simple_metadata(metadata)
return response
@classmethod
def convert_stream_full_response(cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_full_response(
cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream full response.
:param stream_response: stream response
@@ -63,14 +64,14 @@ class ChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, ErrorStreamResponse):
@@ -81,8 +82,9 @@ class ChatAppGenerateResponseConverter(AppGenerateResponseConverter):
yield json.dumps(response_chunk)
@classmethod
def convert_stream_simple_response(cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[ChatbotAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream simple response.
:param stream_response: stream response
@@ -93,20 +95,20 @@ class ChatAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'conversation_id': chunk.conversation_id,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"conversation_id": chunk.conversation_id,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, MessageEndStreamResponse):
sub_stream_response_dict = sub_stream_response.to_dict()
metadata = sub_stream_response_dict.get('metadata', {})
sub_stream_response_dict['metadata'] = cls._get_simple_metadata(metadata)
metadata = sub_stream_response_dict.get("metadata", {})
sub_stream_response_dict["metadata"] = cls._get_simple_metadata(metadata)
response_chunk.update(sub_stream_response_dict)
if isinstance(sub_stream_response, ErrorStreamResponse):
data = cls._error_to_stream_response(sub_stream_response.err)

35
api/core/app/apps/completion/app_config_manager.py
View File

@@ -17,14 +17,15 @@ class CompletionAppConfig(EasyUIBasedAppConfig):
"""
Completion App Config Entity.
"""
pass
class CompletionAppConfigManager(BaseAppConfigManager):
@classmethod
def get_app_config(cls, app_model: App,
app_model_config: AppModelConfig,
override_config_dict: Optional[dict] = None) -> CompletionAppConfig:
def get_app_config(
cls, app_model: App, app_model_config: AppModelConfig, override_config_dict: Optional[dict] = None
) -> CompletionAppConfig:
"""
Convert app model config to completion app config
:param app_model: app model
@@ -51,19 +52,11 @@ class CompletionAppConfigManager(BaseAppConfigManager):
app_model_config_from=config_from,
app_model_config_id=app_model_config.id,
app_model_config_dict=config_dict,
model=ModelConfigManager.convert(
config=config_dict
),
prompt_template=PromptTemplateConfigManager.convert(
config=config_dict
),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(
config=config_dict
),
dataset=DatasetConfigManager.convert(
config=config_dict
),
additional_features=cls.convert_features(config_dict, app_mode)
model=ModelConfigManager.convert(config=config_dict),
prompt_template=PromptTemplateConfigManager.convert(config=config_dict),
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(config=config_dict),
dataset=DatasetConfigManager.convert(config=config_dict),
additional_features=cls.convert_features(config_dict, app_mode),
)
app_config.variables, app_config.external_data_variables = BasicVariablesConfigManager.convert(
@@ -101,8 +94,9 @@ class CompletionAppConfigManager(BaseAppConfigManager):
related_config_keys.extend(current_related_config_keys)
# dataset_query_variable
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(tenant_id, app_mode,
config)
config, current_related_config_keys = DatasetConfigManager.validate_and_set_defaults(
tenant_id, app_mode, config
)
related_config_keys.extend(current_related_config_keys)
# text_to_speech
@@ -114,8 +108,9 @@ class CompletionAppConfigManager(BaseAppConfigManager):
related_config_keys.extend(current_related_config_keys)
# moderation validation
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(tenant_id,
config)
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(
tenant_id, config
)
related_config_keys.extend(current_related_config_keys)
related_config_keys = list(set(related_config_keys))

175
api/core/app/apps/completion/app_generator.py
View File

@@ -32,7 +32,8 @@ logger = logging.getLogger(__name__)
class CompletionAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
@@ -41,19 +42,17 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: Literal[False] = False,
) -> dict: ...
def generate(self, app_model: App,
user: Union[Account, EndUser],
args: Any,
invoke_from: InvokeFrom,
stream: bool = True) \
-> Union[dict, Generator[str, None, None]]:
def generate(
self, app_model: App, user: Union[Account, EndUser], args: Any, invoke_from: InvokeFrom, stream: bool = True
) -> Union[dict, Generator[str, None, None]]:
"""
Generate App response.
@@ -63,12 +62,12 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
:param invoke_from: invoke from source
:param stream: is stream
"""
query = args['query']
query = args["query"]
if not isinstance(query, str):
raise ValueError('query must be a string')
raise ValueError("query must be a string")
query = query.replace('\x00', '')
inputs = args['inputs']
query = query.replace("\x00", "")
inputs = args["inputs"]
extras = {}
@@ -76,41 +75,31 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
conversation = None
# get app model config
app_model_config = self._get_app_model_config(
app_model=app_model,
conversation=conversation
)
app_model_config = self._get_app_model_config(app_model=app_model, conversation=conversation)
# validate override model config
override_model_config_dict = None
if args.get('model_config'):
if args.get("model_config"):
if invoke_from != InvokeFrom.DEBUGGER:
raise ValueError('Only in App debug mode can override model config')
raise ValueError("Only in App debug mode can override model config")
# validate config
override_model_config_dict = CompletionAppConfigManager.config_validate(
tenant_id=app_model.tenant_id,
config=args.get('model_config')
tenant_id=app_model.tenant_id, config=args.get("model_config")
)
# parse files
files = args['files'] if args.get('files') else []
files = args["files"] if args.get("files") else []
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(files, file_extra_config, user)
else:
file_objs = []
# convert to app config
app_config = CompletionAppConfigManager.get_app_config(
app_model=app_model,
app_model_config=app_model_config,
override_config_dict=override_model_config_dict
app_model=app_model, app_model_config=app_model_config, override_config_dict=override_model_config_dict
)
# get tracing instance
@@ -128,14 +117,11 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
stream=stream,
invoke_from=invoke_from,
extras=extras,
trace_manager=trace_manager
trace_manager=trace_manager,
)
# init generate records
(
conversation,
message
) = self._init_generate_records(application_generate_entity)
(conversation, message) = self._init_generate_records(application_generate_entity)
# init queue manager
queue_manager = MessageBasedAppQueueManager(
@@ -144,16 +130,19 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
invoke_from=application_generate_entity.invoke_from,
conversation_id=conversation.id,
app_mode=conversation.mode,
message_id=message.id
message_id=message.id,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(),
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'message_id': message.id,
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(),
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"message_id": message.id,
},
)
worker_thread.start()
@@ -167,15 +156,15 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
stream=stream,
)
return CompletionAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return CompletionAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
def _generate_worker(self, flask_app: Flask,
application_generate_entity: CompletionAppGenerateEntity,
queue_manager: AppQueueManager,
message_id: str) -> None:
def _generate_worker(
self,
flask_app: Flask,
application_generate_entity: CompletionAppGenerateEntity,
queue_manager: AppQueueManager,
message_id: str,
) -> None:
"""
Generate worker in a new thread.
:param flask_app: Flask app
@@ -194,20 +183,19 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
runner.run(
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
message=message
message=message,
)
except GenerateTaskStoppedException:
pass
except InvokeAuthorizationError:
queue_manager.publish_error(
InvokeAuthorizationError('Incorrect API key provided'),
PublishFrom.APPLICATION_MANAGER
InvokeAuthorizationError("Incorrect API key provided"), PublishFrom.APPLICATION_MANAGER
)
except ValidationError as e:
logger.exception("Validation Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except (ValueError, InvokeError) as e:
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == 'true':
if os.environ.get("DEBUG") and os.environ.get("DEBUG").lower() == "true":
logger.exception("Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except Exception as e:
@@ -216,12 +204,14 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
finally:
db.session.close()
def generate_more_like_this(self, app_model: App,
message_id: str,
user: Union[Account, EndUser],
invoke_from: InvokeFrom,
stream: bool = True) \
-> Union[dict, Generator[str, None, None]]:
def generate_more_like_this(
self,
app_model: App,
message_id: str,
user: Union[Account, EndUser],
invoke_from: InvokeFrom,
stream: bool = True,
) -> Union[dict, Generator[str, None, None]]:
"""
Generate App response.
@@ -231,13 +221,17 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
:param invoke_from: invoke from source
:param stream: is stream
"""
message = db.session.query(Message).filter(
Message.id == message_id,
Message.app_id == app_model.id,
Message.from_source == ('api' if isinstance(user, EndUser) else 'console'),
Message.from_end_user_id == (user.id if isinstance(user, EndUser) else None),
Message.from_account_id == (user.id if isinstance(user, Account) else None),
).first()
message = (
db.session.query(Message)
.filter(
Message.id == message_id,
Message.app_id == app_model.id,
Message.from_source == ("api" if isinstance(user, EndUser) else "console"),
Message.from_end_user_id == (user.id if isinstance(user, EndUser) else None),
Message.from_account_id == (user.id if isinstance(user, Account) else None),
)
.first()
)
if not message:
raise MessageNotExistsError()
@@ -250,29 +244,23 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
app_model_config = message.app_model_config
override_model_config_dict = app_model_config.to_dict()
model_dict = override_model_config_dict['model']
completion_params = model_dict.get('completion_params')
completion_params['temperature'] = 0.9
model_dict['completion_params'] = completion_params
override_model_config_dict['model'] = model_dict
model_dict = override_model_config_dict["model"]
completion_params = model_dict.get("completion_params")
completion_params["temperature"] = 0.9
model_dict["completion_params"] = completion_params
override_model_config_dict["model"] = model_dict
# parse files
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(override_model_config_dict or app_model_config.to_dict())
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
message.files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(message.files, file_extra_config, user)
else:
file_objs = []
# convert to app config
app_config = CompletionAppConfigManager.get_app_config(
app_model=app_model,
app_model_config=app_model_config,
override_config_dict=override_model_config_dict
app_model=app_model, app_model_config=app_model_config, override_config_dict=override_model_config_dict
)
# init application generate entity
@@ -286,14 +274,11 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
user_id=user.id,
stream=stream,
invoke_from=invoke_from,
extras={}
extras={},
)
# init generate records
(
conversation,
message
) = self._init_generate_records(application_generate_entity)
(conversation, message) = self._init_generate_records(application_generate_entity)
# init queue manager
queue_manager = MessageBasedAppQueueManager(
@@ -302,16 +287,19 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
invoke_from=application_generate_entity.invoke_from,
conversation_id=conversation.id,
app_mode=conversation.mode,
message_id=message.id
message_id=message.id,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(),
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'message_id': message.id,
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(),
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"message_id": message.id,
},
)
worker_thread.start()
@@ -325,7 +313,4 @@ class CompletionAppGenerator(MessageBasedAppGenerator):
stream=stream,
)
return CompletionAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return CompletionAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)

36
api/core/app/apps/completion/app_runner.py
View File

@@ -22,9 +22,9 @@ class CompletionAppRunner(AppRunner):
Completion Application Runner
"""
def run(self, application_generate_entity: CompletionAppGenerateEntity,
queue_manager: AppQueueManager,
message: Message) -> None:
def run(
self, application_generate_entity: CompletionAppGenerateEntity, queue_manager: AppQueueManager, message: Message
) -> None:
"""
Run application
:param application_generate_entity: application generate entity
@@ -54,7 +54,7 @@ class CompletionAppRunner(AppRunner):
prompt_template_entity=app_config.prompt_template,
inputs=inputs,
files=files,
query=query
query=query,
)
# organize all inputs and template to prompt messages
@@ -65,7 +65,7 @@ class CompletionAppRunner(AppRunner):
prompt_template_entity=app_config.prompt_template,
inputs=inputs,
files=files,
query=query
query=query,
)
# moderation
@@ -77,7 +77,7 @@ class CompletionAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
inputs=inputs,
query=query,
message_id=message.id
message_id=message.id,
)
except ModerationException as e:
self.direct_output(
@@ -85,7 +85,7 @@ class CompletionAppRunner(AppRunner):
app_generate_entity=application_generate_entity,
prompt_messages=prompt_messages,
text=str(e),
stream=application_generate_entity.stream
stream=application_generate_entity.stream,
)
return
@@ -97,7 +97,7 @@ class CompletionAppRunner(AppRunner):
app_id=app_record.id,
external_data_tools=external_data_tools,
inputs=inputs,
query=query
query=query,
)
# get context from datasets
@@ -108,7 +108,7 @@ class CompletionAppRunner(AppRunner):
app_record.id,
message.id,
application_generate_entity.user_id,
application_generate_entity.invoke_from
application_generate_entity.invoke_from,
)
dataset_config = app_config.dataset
@@ -126,7 +126,7 @@ class CompletionAppRunner(AppRunner):
invoke_from=application_generate_entity.invoke_from,
show_retrieve_source=app_config.additional_features.show_retrieve_source,
hit_callback=hit_callback,
message_id=message.id
message_id=message.id,
)
# reorganize all inputs and template to prompt messages
@@ -139,29 +139,26 @@ class CompletionAppRunner(AppRunner):
inputs=inputs,
files=files,
query=query,
context=context
context=context,
)
# check hosting moderation
hosting_moderation_result = self.check_hosting_moderation(
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
prompt_messages=prompt_messages
prompt_messages=prompt_messages,
)
if hosting_moderation_result:
return
# Re-calculate the max tokens if sum(prompt_token + max_tokens) over model token limit
self.recalc_llm_max_tokens(
model_config=application_generate_entity.model_conf,
prompt_messages=prompt_messages
)
self.recalc_llm_max_tokens(model_config=application_generate_entity.model_conf, prompt_messages=prompt_messages)
# Invoke model
model_instance = ModelInstance(
provider_model_bundle=application_generate_entity.model_conf.provider_model_bundle,
model=application_generate_entity.model_conf.model
model=application_generate_entity.model_conf.model,
)
db.session.close()
@@ -176,8 +173,5 @@ class CompletionAppRunner(AppRunner):
# handle invoke result
self._handle_invoke_result(
invoke_result=invoke_result,
queue_manager=queue_manager,
stream=application_generate_entity.stream
invoke_result=invoke_result, queue_manager=queue_manager, stream=application_generate_entity.stream
)

50
api/core/app/apps/completion/generate_response_converter.py
View File

@@ -23,14 +23,14 @@ class CompletionAppGenerateResponseConverter(AppGenerateResponseConverter):
:return:
"""
response = {
'event': 'message',
'task_id': blocking_response.task_id,
'id': blocking_response.data.id,
'message_id': blocking_response.data.message_id,
'mode': blocking_response.data.mode,
'answer': blocking_response.data.answer,
'metadata': blocking_response.data.metadata,
'created_at': blocking_response.data.created_at
"event": "message",
"task_id": blocking_response.task_id,
"id": blocking_response.data.id,
"message_id": blocking_response.data.message_id,
"mode": blocking_response.data.mode,
"answer": blocking_response.data.answer,
"metadata": blocking_response.data.metadata,
"created_at": blocking_response.data.created_at,
}
return response
@@ -44,14 +44,15 @@ class CompletionAppGenerateResponseConverter(AppGenerateResponseConverter):
"""
response = cls.convert_blocking_full_response(blocking_response)
metadata = response.get('metadata', {})
response['metadata'] = cls._get_simple_metadata(metadata)
metadata = response.get("metadata", {})
response["metadata"] = cls._get_simple_metadata(metadata)
return response
@classmethod
def convert_stream_full_response(cls, stream_response: Generator[CompletionAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_full_response(
cls, stream_response: Generator[CompletionAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream full response.
:param stream_response: stream response
@@ -62,13 +63,13 @@ class CompletionAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, ErrorStreamResponse):
@@ -79,8 +80,9 @@ class CompletionAppGenerateResponseConverter(AppGenerateResponseConverter):
yield json.dumps(response_chunk)
@classmethod
def convert_stream_simple_response(cls, stream_response: Generator[CompletionAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[CompletionAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream simple response.
:param stream_response: stream response
@@ -91,19 +93,19 @@ class CompletionAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'message_id': chunk.message_id,
'created_at': chunk.created_at
"event": sub_stream_response.event.value,
"message_id": chunk.message_id,
"created_at": chunk.created_at,
}
if isinstance(sub_stream_response, MessageEndStreamResponse):
sub_stream_response_dict = sub_stream_response.to_dict()
metadata = sub_stream_response_dict.get('metadata', {})
sub_stream_response_dict['metadata'] = cls._get_simple_metadata(metadata)
metadata = sub_stream_response_dict.get("metadata", {})
sub_stream_response_dict["metadata"] = cls._get_simple_metadata(metadata)
response_chunk.update(sub_stream_response_dict)
if isinstance(sub_stream_response, ErrorStreamResponse):
data = cls._error_to_stream_response(sub_stream_response.err)

104
api/core/app/apps/message_based_app_generator.py
View File

@@ -35,23 +35,23 @@ logger = logging.getLogger(__name__)
class MessageBasedAppGenerator(BaseAppGenerator):
def _handle_response(
self, application_generate_entity: Union[
ChatAppGenerateEntity,
CompletionAppGenerateEntity,
AgentChatAppGenerateEntity,
AdvancedChatAppGenerateEntity
],
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
user: Union[Account, EndUser],
stream: bool = False,
self,
application_generate_entity: Union[
ChatAppGenerateEntity,
CompletionAppGenerateEntity,
AgentChatAppGenerateEntity,
AdvancedChatAppGenerateEntity,
],
queue_manager: AppQueueManager,
conversation: Conversation,
message: Message,
user: Union[Account, EndUser],
stream: bool = False,
) -> Union[
ChatbotAppBlockingResponse,
CompletionAppBlockingResponse,
Generator[Union[ChatbotAppStreamResponse, CompletionAppStreamResponse], None, None]
Generator[Union[ChatbotAppStreamResponse, CompletionAppStreamResponse], None, None],
]:
"""
Handle response.
@@ -70,7 +70,7 @@ class MessageBasedAppGenerator(BaseAppGenerator):
conversation=conversation,
message=message,
user=user,
stream=stream
stream=stream,
)
try:
@@ -82,12 +82,13 @@ class MessageBasedAppGenerator(BaseAppGenerator):
logger.exception(e)
raise e
def _get_conversation_by_user(self, app_model: App, conversation_id: str,
user: Union[Account, EndUser]) -> Conversation:
def _get_conversation_by_user(
self, app_model: App, conversation_id: str, user: Union[Account, EndUser]
) -> Conversation:
conversation_filter = [
Conversation.id == conversation_id,
Conversation.app_id == app_model.id,
Conversation.status == 'normal'
Conversation.status == "normal",
]
if isinstance(user, Account):
@@ -100,19 +101,18 @@ class MessageBasedAppGenerator(BaseAppGenerator):
if not conversation:
raise ConversationNotExistsError()
if conversation.status != 'normal':
if conversation.status != "normal":
raise ConversationCompletedError()
return conversation
def _get_app_model_config(self, app_model: App,
conversation: Optional[Conversation] = None) \
-> AppModelConfig:
def _get_app_model_config(self, app_model: App, conversation: Optional[Conversation] = None) -> AppModelConfig:
if conversation:
app_model_config = db.session.query(AppModelConfig).filter(
AppModelConfig.id == conversation.app_model_config_id,
AppModelConfig.app_id == app_model.id
).first()
app_model_config = (
db.session.query(AppModelConfig)
.filter(AppModelConfig.id == conversation.app_model_config_id, AppModelConfig.app_id == app_model.id)
.first()
)
if not app_model_config:
raise AppModelConfigBrokenError()
@@ -127,15 +127,16 @@ class MessageBasedAppGenerator(BaseAppGenerator):
return app_model_config
def _init_generate_records(self,
application_generate_entity: Union[
ChatAppGenerateEntity,
CompletionAppGenerateEntity,
AgentChatAppGenerateEntity,
AdvancedChatAppGenerateEntity
],
conversation: Optional[Conversation] = None) \
-> tuple[Conversation, Message]:
def _init_generate_records(
self,
application_generate_entity: Union[
ChatAppGenerateEntity,
CompletionAppGenerateEntity,
AgentChatAppGenerateEntity,
AdvancedChatAppGenerateEntity,
],
conversation: Optional[Conversation] = None,
) -> tuple[Conversation, Message]:
"""
Initialize generate records
:param application_generate_entity: application generate entity
@@ -148,10 +149,10 @@ class MessageBasedAppGenerator(BaseAppGenerator):
end_user_id = None
account_id = None
if application_generate_entity.invoke_from in [InvokeFrom.WEB_APP, InvokeFrom.SERVICE_API]:
from_source = 'api'
from_source = "api"
end_user_id = application_generate_entity.user_id
else:
from_source = 'console'
from_source = "console"
account_id = application_generate_entity.user_id
if isinstance(application_generate_entity, AdvancedChatAppGenerateEntity):
@@ -164,8 +165,11 @@ class MessageBasedAppGenerator(BaseAppGenerator):
model_provider = application_generate_entity.model_conf.provider
model_id = application_generate_entity.model_conf.model
override_model_configs = None
if app_config.app_model_config_from == EasyUIBasedAppModelConfigFrom.ARGS \
and app_config.app_mode in [AppMode.AGENT_CHAT, AppMode.CHAT, AppMode.COMPLETION]:
if app_config.app_model_config_from == EasyUIBasedAppModelConfigFrom.ARGS and app_config.app_mode in [
AppMode.AGENT_CHAT,
AppMode.CHAT,
AppMode.COMPLETION,
]:
override_model_configs = app_config.app_model_config_dict
# get conversation introduction
@@ -179,12 +183,12 @@ class MessageBasedAppGenerator(BaseAppGenerator):
model_id=model_id,
override_model_configs=json.dumps(override_model_configs) if override_model_configs else None,
mode=app_config.app_mode.value,
name='New conversation',
name="New conversation",
inputs=application_generate_entity.inputs,
introduction=introduction,
system_instruction="",
system_instruction_tokens=0,
status='normal',
status="normal",
invoke_from=application_generate_entity.invoke_from.value,
from_source=from_source,
from_end_user_id=end_user_id,
@@ -216,11 +220,11 @@ class MessageBasedAppGenerator(BaseAppGenerator):
answer_price_unit=0,
provider_response_latency=0,
total_price=0,
currency='USD',
currency="USD",
invoke_from=application_generate_entity.invoke_from.value,
from_source=from_source,
from_end_user_id=end_user_id,
from_account_id=account_id
from_account_id=account_id,
)
db.session.add(message)
@@ -232,10 +236,10 @@ class MessageBasedAppGenerator(BaseAppGenerator):
message_id=message.id,
type=file.type.value,
transfer_method=file.transfer_method.value,
belongs_to='user',
belongs_to="user",
url=file.url,
upload_file_id=file.related_id,
created_by_role=('account' if account_id else 'end_user'),
created_by_role=("account" if account_id else "end_user"),
created_by=account_id or end_user_id,
)
db.session.add(message_file)
@@ -269,11 +273,7 @@ class MessageBasedAppGenerator(BaseAppGenerator):
:param conversation_id: conversation id
:return: conversation
"""
conversation = (
db.session.query(Conversation)
.filter(Conversation.id == conversation_id)
.first()
)
conversation = db.session.query(Conversation).filter(Conversation.id == conversation_id).first()
if not conversation:
raise ConversationNotExistsError()
@@ -286,10 +286,6 @@ class MessageBasedAppGenerator(BaseAppGenerator):
:param message_id: message id
:return: message
"""
message = (
db.session.query(Message)
.filter(Message.id == message_id)
.first()
)
message = db.session.query(Message).filter(Message.id == message_id).first()
return message

21
api/core/app/apps/message_based_app_queue_manager.py
View File

@@ -12,12 +12,9 @@ from core.app.entities.queue_entities import (
class MessageBasedAppQueueManager(AppQueueManager):
def __init__(self, task_id: str,
user_id: str,
invoke_from: InvokeFrom,
conversation_id: str,
app_mode: str,
message_id: str) -> None:
def __init__(
self, task_id: str, user_id: str, invoke_from: InvokeFrom, conversation_id: str, app_mode: str, message_id: str
) -> None:
super().__init__(task_id, user_id, invoke_from)
self._conversation_id = str(conversation_id)
@@ -30,7 +27,7 @@ class MessageBasedAppQueueManager(AppQueueManager):
message_id=self._message_id,
conversation_id=self._conversation_id,
app_mode=self._app_mode,
event=event
event=event,
)
def _publish(self, event: AppQueueEvent, pub_from: PublishFrom) -> None:
@@ -45,17 +42,15 @@ class MessageBasedAppQueueManager(AppQueueManager):
message_id=self._message_id,
conversation_id=self._conversation_id,
app_mode=self._app_mode,
event=event
event=event,
)
self._q.put(message)
if isinstance(event, QueueStopEvent
| QueueErrorEvent
| QueueMessageEndEvent
| QueueAdvancedChatMessageEndEvent):
if isinstance(
event, QueueStopEvent | QueueErrorEvent | QueueMessageEndEvent | QueueAdvancedChatMessageEndEvent
):
self.stop_listen()
if pub_from == PublishFrom.APPLICATION_MANAGER and self._is_stopped():
raise GenerateTaskStoppedException()

18
api/core/app/apps/workflow/app_config_manager.py
View File

@@ -12,6 +12,7 @@ class WorkflowAppConfig(WorkflowUIBasedAppConfig):
"""
Workflow App Config Entity.
"""
pass
@@ -26,13 +27,9 @@ class WorkflowAppConfigManager(BaseAppConfigManager):
app_id=app_model.id,
app_mode=app_mode,
workflow_id=workflow.id,
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(
config=features_dict
),
variables=WorkflowVariablesConfigManager.convert(
workflow=workflow
),
additional_features=cls.convert_features(features_dict, app_mode)
sensitive_word_avoidance=SensitiveWordAvoidanceConfigManager.convert(config=features_dict),
variables=WorkflowVariablesConfigManager.convert(workflow=workflow),
additional_features=cls.convert_features(features_dict, app_mode),
)
return app_config
@@ -50,8 +47,7 @@ class WorkflowAppConfigManager(BaseAppConfigManager):
# file upload validation
config, current_related_config_keys = FileUploadConfigManager.validate_and_set_defaults(
config=config,
is_vision=False
config=config, is_vision=False
)
related_config_keys.extend(current_related_config_keys)
@@ -61,9 +57,7 @@ class WorkflowAppConfigManager(BaseAppConfigManager):
# moderation validation
config, current_related_config_keys = SensitiveWordAvoidanceConfigManager.validate_and_set_defaults(
tenant_id=tenant_id,
config=config,
only_structure_validate=only_structure_validate
tenant_id=tenant_id, config=config, only_structure_validate=only_structure_validate
)
related_config_keys.extend(current_related_config_keys)

129
api/core/app/apps/workflow/app_generator.py
View File

@@ -34,26 +34,28 @@ logger = logging.getLogger(__name__)
class WorkflowAppGenerator(BaseAppGenerator):
@overload
def generate(
self, app_model: App,
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: Literal[True] = True,
call_depth: int = 0,
workflow_thread_pool_id: Optional[str] = None
workflow_thread_pool_id: Optional[str] = None,
) -> Generator[str, None, None]: ...
@overload
def generate(
self, app_model: App,
self,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
args: dict,
invoke_from: InvokeFrom,
stream: Literal[False] = False,
call_depth: int = 0,
workflow_thread_pool_id: Optional[str] = None
workflow_thread_pool_id: Optional[str] = None,
) -> dict: ...
def generate(
@@ -65,7 +67,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
invoke_from: InvokeFrom,
stream: bool = True,
call_depth: int = 0,
workflow_thread_pool_id: Optional[str] = None
workflow_thread_pool_id: Optional[str] = None,
):
"""
Generate App response.
@@ -79,26 +81,19 @@ class WorkflowAppGenerator(BaseAppGenerator):
:param call_depth: call depth
:param workflow_thread_pool_id: workflow thread pool id
"""
inputs = args['inputs']
inputs = args["inputs"]
# parse files
files = args['files'] if args.get('files') else []
files = args["files"] if args.get("files") else []
message_file_parser = MessageFileParser(tenant_id=app_model.tenant_id, app_id=app_model.id)
file_extra_config = FileUploadConfigManager.convert(workflow.features_dict, is_vision=False)
if file_extra_config:
file_objs = message_file_parser.validate_and_transform_files_arg(
files,
file_extra_config,
user
)
file_objs = message_file_parser.validate_and_transform_files_arg(files, file_extra_config, user)
else:
file_objs = []
# convert to app config
app_config = WorkflowAppConfigManager.get_app_config(
app_model=app_model,
workflow=workflow
)
app_config = WorkflowAppConfigManager.get_app_config(app_model=app_model, workflow=workflow)
# get tracing instance
user_id = user.id if isinstance(user, Account) else user.session_id
@@ -114,7 +109,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
stream=stream,
invoke_from=invoke_from,
call_depth=call_depth,
trace_manager=trace_manager
trace_manager=trace_manager,
)
contexts.tenant_id.set(application_generate_entity.app_config.tenant_id)
@@ -125,18 +120,19 @@ class WorkflowAppGenerator(BaseAppGenerator):
application_generate_entity=application_generate_entity,
invoke_from=invoke_from,
stream=stream,
workflow_thread_pool_id=workflow_thread_pool_id
workflow_thread_pool_id=workflow_thread_pool_id,
)
def _generate(
self, *,
self,
*,
app_model: App,
workflow: Workflow,
user: Union[Account, EndUser],
application_generate_entity: WorkflowAppGenerateEntity,
invoke_from: InvokeFrom,
stream: bool = True,
workflow_thread_pool_id: Optional[str] = None
workflow_thread_pool_id: Optional[str] = None,
) -> dict[str, Any] | Generator[str, None, None]:
"""
Generate App response.
@@ -154,17 +150,20 @@ class WorkflowAppGenerator(BaseAppGenerator):
task_id=application_generate_entity.task_id,
user_id=application_generate_entity.user_id,
invoke_from=application_generate_entity.invoke_from,
app_mode=app_model.mode
app_mode=app_model.mode,
)
# new thread
worker_thread = threading.Thread(target=self._generate_worker, kwargs={
'flask_app': current_app._get_current_object(), # type: ignore
'application_generate_entity': application_generate_entity,
'queue_manager': queue_manager,
'context': contextvars.copy_context(),
'workflow_thread_pool_id': workflow_thread_pool_id
})
worker_thread = threading.Thread(
target=self._generate_worker,
kwargs={
"flask_app": current_app._get_current_object(), # type: ignore
"application_generate_entity": application_generate_entity,
"queue_manager": queue_manager,
"context": contextvars.copy_context(),
"workflow_thread_pool_id": workflow_thread_pool_id,
},
)
worker_thread.start()
@@ -177,17 +176,11 @@ class WorkflowAppGenerator(BaseAppGenerator):
stream=stream,
)
return WorkflowAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from
)
return WorkflowAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
def single_iteration_generate(self, app_model: App,
workflow: Workflow,
node_id: str,
user: Account,
args: dict,
stream: bool = True) -> dict[str, Any] | Generator[str, Any, None]:
def single_iteration_generate(
self, app_model: App, workflow: Workflow, node_id: str, user: Account, args: dict, stream: bool = True
) -> dict[str, Any] | Generator[str, Any, None]:
"""
Generate App response.
@@ -199,16 +192,13 @@ class WorkflowAppGenerator(BaseAppGenerator):
:param stream: is stream
"""
if not node_id:
raise ValueError('node_id is required')
raise ValueError("node_id is required")
if args.get('inputs') is None:
raise ValueError('inputs is required')
if args.get("inputs") is None:
raise ValueError("inputs is required")
# convert to app config
app_config = WorkflowAppConfigManager.get_app_config(
app_model=app_model,
workflow=workflow
)
app_config = WorkflowAppConfigManager.get_app_config(app_model=app_model, workflow=workflow)
# init application generate entity
application_generate_entity = WorkflowAppGenerateEntity(
@@ -219,13 +209,10 @@ class WorkflowAppGenerator(BaseAppGenerator):
user_id=user.id,
stream=stream,
invoke_from=InvokeFrom.DEBUGGER,
extras={
"auto_generate_conversation_name": False
},
extras={"auto_generate_conversation_name": False},
single_iteration_run=WorkflowAppGenerateEntity.SingleIterationRunEntity(
node_id=node_id,
inputs=args['inputs']
)
node_id=node_id, inputs=args["inputs"]
),
)
contexts.tenant_id.set(application_generate_entity.app_config.tenant_id)
@@ -235,14 +222,17 @@ class WorkflowAppGenerator(BaseAppGenerator):
user=user,
invoke_from=InvokeFrom.DEBUGGER,
application_generate_entity=application_generate_entity,
stream=stream
stream=stream,
)
def _generate_worker(self, flask_app: Flask,
application_generate_entity: WorkflowAppGenerateEntity,
queue_manager: AppQueueManager,
context: contextvars.Context,
workflow_thread_pool_id: Optional[str] = None) -> None:
def _generate_worker(
self,
flask_app: Flask,
application_generate_entity: WorkflowAppGenerateEntity,
queue_manager: AppQueueManager,
context: contextvars.Context,
workflow_thread_pool_id: Optional[str] = None,
) -> None:
"""
Generate worker in a new thread.
:param flask_app: Flask app
@@ -259,7 +249,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
runner = WorkflowAppRunner(
application_generate_entity=application_generate_entity,
queue_manager=queue_manager,
workflow_thread_pool_id=workflow_thread_pool_id
workflow_thread_pool_id=workflow_thread_pool_id,
)
runner.run()
@@ -267,14 +257,13 @@ class WorkflowAppGenerator(BaseAppGenerator):
pass
except InvokeAuthorizationError:
queue_manager.publish_error(
InvokeAuthorizationError('Incorrect API key provided'),
PublishFrom.APPLICATION_MANAGER
InvokeAuthorizationError("Incorrect API key provided"), PublishFrom.APPLICATION_MANAGER
)
except ValidationError as e:
logger.exception("Validation Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except (ValueError, InvokeError) as e:
if os.environ.get("DEBUG") and os.environ.get("DEBUG", "false").lower() == 'true':
if os.environ.get("DEBUG") and os.environ.get("DEBUG", "false").lower() == "true":
logger.exception("Error when generating")
queue_manager.publish_error(e, PublishFrom.APPLICATION_MANAGER)
except Exception as e:
@@ -283,14 +272,14 @@ class WorkflowAppGenerator(BaseAppGenerator):
finally:
db.session.close()
def _handle_response(self, application_generate_entity: WorkflowAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
user: Union[Account, EndUser],
stream: bool = False) -> Union[
WorkflowAppBlockingResponse,
Generator[WorkflowAppStreamResponse, None, None]
]:
def _handle_response(
self,
application_generate_entity: WorkflowAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
user: Union[Account, EndUser],
stream: bool = False,
) -> Union[WorkflowAppBlockingResponse, Generator[WorkflowAppStreamResponse, None, None]]:
"""
Handle response.
:param application_generate_entity: application generate entity
@@ -306,7 +295,7 @@ class WorkflowAppGenerator(BaseAppGenerator):
workflow=workflow,
queue_manager=queue_manager,
user=user,
stream=stream
stream=stream,
)
try:

24
api/core/app/apps/workflow/app_queue_manager.py
View File

@@ -12,10 +12,7 @@ from core.app.entities.queue_entities import (
class WorkflowAppQueueManager(AppQueueManager):
def __init__(self, task_id: str,
user_id: str,
invoke_from: InvokeFrom,
app_mode: str) -> None:
def __init__(self, task_id: str, user_id: str, invoke_from: InvokeFrom, app_mode: str) -> None:
super().__init__(task_id, user_id, invoke_from)
self._app_mode = app_mode
@@ -27,19 +24,18 @@ class WorkflowAppQueueManager(AppQueueManager):
:param pub_from:
:return:
"""
message = WorkflowQueueMessage(
task_id=self._task_id,
app_mode=self._app_mode,
event=event
)
message = WorkflowQueueMessage(task_id=self._task_id, app_mode=self._app_mode, event=event)
self._q.put(message)
if isinstance(event, QueueStopEvent
| QueueErrorEvent
| QueueMessageEndEvent
| QueueWorkflowSucceededEvent
| QueueWorkflowFailedEvent):
if isinstance(
event,
QueueStopEvent
| QueueErrorEvent
| QueueMessageEndEvent
| QueueWorkflowSucceededEvent
| QueueWorkflowFailedEvent,
):
self.stop_listen()
if pub_from == PublishFrom.APPLICATION_MANAGER and self._is_stopped():

23
api/core/app/apps/workflow/app_runner.py
View File

@@ -28,10 +28,10 @@ class WorkflowAppRunner(WorkflowBasedAppRunner):
"""
def __init__(
self,
application_generate_entity: WorkflowAppGenerateEntity,
queue_manager: AppQueueManager,
workflow_thread_pool_id: Optional[str] = None
self,
application_generate_entity: WorkflowAppGenerateEntity,
queue_manager: AppQueueManager,
workflow_thread_pool_id: Optional[str] = None,
) -> None:
"""
:param application_generate_entity: application generate entity
@@ -62,16 +62,16 @@ class WorkflowAppRunner(WorkflowBasedAppRunner):
app_record = db.session.query(App).filter(App.id == app_config.app_id).first()
if not app_record:
raise ValueError('App not found')
raise ValueError("App not found")
workflow = self.get_workflow(app_model=app_record, workflow_id=app_config.workflow_id)
if not workflow:
raise ValueError('Workflow not initialized')
raise ValueError("Workflow not initialized")
db.session.close()
workflow_callbacks: list[WorkflowCallback] = []
if bool(os.environ.get('DEBUG', 'False').lower() == 'true'):
if bool(os.environ.get("DEBUG", "False").lower() == "true"):
workflow_callbacks.append(WorkflowLoggingCallback())
# if only single iteration run is requested
@@ -80,10 +80,9 @@ class WorkflowAppRunner(WorkflowBasedAppRunner):
graph, variable_pool = self._get_graph_and_variable_pool_of_single_iteration(
workflow=workflow,
node_id=self.application_generate_entity.single_iteration_run.node_id,
user_inputs=self.application_generate_entity.single_iteration_run.inputs
user_inputs=self.application_generate_entity.single_iteration_run.inputs,
)
else:
inputs = self.application_generate_entity.inputs
files = self.application_generate_entity.files
@@ -120,12 +119,10 @@ class WorkflowAppRunner(WorkflowBasedAppRunner):
invoke_from=self.application_generate_entity.invoke_from,
call_depth=self.application_generate_entity.call_depth,
variable_pool=variable_pool,
thread_pool_id=self.workflow_thread_pool_id
thread_pool_id=self.workflow_thread_pool_id,
)
generator = workflow_entry.run(
callbacks=workflow_callbacks
)
generator = workflow_entry.run(callbacks=workflow_callbacks)
for event in generator:
self._handle_event(workflow_entry, event)

22
api/core/app/apps/workflow/generate_response_converter.py
View File

@@ -35,8 +35,9 @@ class WorkflowAppGenerateResponseConverter(AppGenerateResponseConverter):
return cls.convert_blocking_full_response(blocking_response)
@classmethod
def convert_stream_full_response(cls, stream_response: Generator[WorkflowAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_full_response(
cls, stream_response: Generator[WorkflowAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream full response.
:param stream_response: stream response
@@ -47,12 +48,12 @@ class WorkflowAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'workflow_run_id': chunk.workflow_run_id,
"event": sub_stream_response.event.value,
"workflow_run_id": chunk.workflow_run_id,
}
if isinstance(sub_stream_response, ErrorStreamResponse):
@@ -63,8 +64,9 @@ class WorkflowAppGenerateResponseConverter(AppGenerateResponseConverter):
yield json.dumps(response_chunk)
@classmethod
def convert_stream_simple_response(cls, stream_response: Generator[WorkflowAppStreamResponse, None, None]) \
-> Generator[str, None, None]:
def convert_stream_simple_response(
cls, stream_response: Generator[WorkflowAppStreamResponse, None, None]
) -> Generator[str, None, None]:
"""
Convert stream simple response.
:param stream_response: stream response
@@ -75,12 +77,12 @@ class WorkflowAppGenerateResponseConverter(AppGenerateResponseConverter):
sub_stream_response = chunk.stream_response
if isinstance(sub_stream_response, PingStreamResponse):
yield 'ping'
yield "ping"
continue
response_chunk = {
'event': sub_stream_response.event.value,
'workflow_run_id': chunk.workflow_run_id,
"event": sub_stream_response.event.value,
"workflow_run_id": chunk.workflow_run_id,
}
if isinstance(sub_stream_response, ErrorStreamResponse):

133
api/core/app/apps/workflow/generate_task_pipeline.py
View File

@@ -63,17 +63,21 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
"""
WorkflowAppGenerateTaskPipeline is a class that generate stream output and state management for Application.
"""
_workflow: Workflow
_user: Union[Account, EndUser]
_task_state: WorkflowTaskState
_application_generate_entity: WorkflowAppGenerateEntity
_workflow_system_variables: dict[SystemVariableKey, Any]
def __init__(self, application_generate_entity: WorkflowAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
user: Union[Account, EndUser],
stream: bool) -> None:
def __init__(
self,
application_generate_entity: WorkflowAppGenerateEntity,
workflow: Workflow,
queue_manager: AppQueueManager,
user: Union[Account, EndUser],
stream: bool,
) -> None:
"""
Initialize GenerateTaskPipeline.
:param application_generate_entity: application generate entity
@@ -92,7 +96,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
self._workflow = workflow
self._workflow_system_variables = {
SystemVariableKey.FILES: application_generate_entity.files,
SystemVariableKey.USER_ID: user_id
SystemVariableKey.USER_ID: user_id,
}
self._task_state = WorkflowTaskState()
@@ -106,16 +110,13 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
db.session.refresh(self._user)
db.session.close()
generator = self._wrapper_process_stream_response(
trace_manager=self._application_generate_entity.trace_manager
)
generator = self._wrapper_process_stream_response(trace_manager=self._application_generate_entity.trace_manager)
if self._stream:
return self._to_stream_response(generator)
else:
return self._to_blocking_response(generator)
def _to_blocking_response(self, generator: Generator[StreamResponse, None, None]) \
-> WorkflowAppBlockingResponse:
def _to_blocking_response(self, generator: Generator[StreamResponse, None, None]) -> WorkflowAppBlockingResponse:
"""
To blocking response.
:return:
@@ -137,18 +138,19 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
total_tokens=stream_response.data.total_tokens,
total_steps=stream_response.data.total_steps,
created_at=int(stream_response.data.created_at),
finished_at=int(stream_response.data.finished_at)
)
finished_at=int(stream_response.data.finished_at),
),
)
return response
else:
continue
raise Exception('Queue listening stopped unexpectedly.')
raise Exception("Queue listening stopped unexpectedly.")
def _to_stream_response(self, generator: Generator[StreamResponse, None, None]) \
-> Generator[WorkflowAppStreamResponse, None, None]:
def _to_stream_response(
self, generator: Generator[StreamResponse, None, None]
) -> Generator[WorkflowAppStreamResponse, None, None]:
"""
To stream response.
:return:
@@ -158,10 +160,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
if isinstance(stream_response, WorkflowStartStreamResponse):
workflow_run_id = stream_response.workflow_run_id
yield WorkflowAppStreamResponse(
workflow_run_id=workflow_run_id,
stream_response=stream_response
)
yield WorkflowAppStreamResponse(workflow_run_id=workflow_run_id, stream_response=stream_response)
def _listenAudioMsg(self, publisher, task_id: str):
if not publisher:
@@ -171,17 +170,20 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
return MessageAudioStreamResponse(audio=audio_msg.audio, task_id=task_id)
return None
def _wrapper_process_stream_response(self, trace_manager: Optional[TraceQueueManager] = None) -> \
Generator[StreamResponse, None, None]:
def _wrapper_process_stream_response(
self, trace_manager: Optional[TraceQueueManager] = None
) -> Generator[StreamResponse, None, None]:
tts_publisher = None
task_id = self._application_generate_entity.task_id
tenant_id = self._application_generate_entity.app_config.tenant_id
features_dict = self._workflow.features_dict
if features_dict.get('text_to_speech') and features_dict['text_to_speech'].get('enabled') and features_dict[
'text_to_speech'].get('autoPlay') == 'enabled':
tts_publisher = AppGeneratorTTSPublisher(tenant_id, features_dict['text_to_speech'].get('voice'))
if (
features_dict.get("text_to_speech")
and features_dict["text_to_speech"].get("enabled")
and features_dict["text_to_speech"].get("autoPlay") == "enabled"
):
tts_publisher = AppGeneratorTTSPublisher(tenant_id, features_dict["text_to_speech"].get("voice"))
for response in self._process_stream_response(tts_publisher=tts_publisher, trace_manager=trace_manager):
while True:
@@ -210,13 +212,12 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
except Exception as e:
logger.error(e)
break
yield MessageAudioEndStreamResponse(audio='', task_id=task_id)
yield MessageAudioEndStreamResponse(audio="", task_id=task_id)
def _process_stream_response(
self,
tts_publisher: Optional[AppGeneratorTTSPublisher] = None,
trace_manager: Optional[TraceQueueManager] = None
trace_manager: Optional[TraceQueueManager] = None,
) -> Generator[StreamResponse, None, None]:
"""
Process stream response.
@@ -241,22 +242,18 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
# init workflow run
workflow_run = self._handle_workflow_run_start()
yield self._workflow_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
elif isinstance(event, QueueNodeStartedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
workflow_node_execution = self._handle_node_execution_start(
workflow_run=workflow_run,
event=event
)
workflow_node_execution = self._handle_node_execution_start(workflow_run=workflow_run, event=event)
response = self._workflow_node_start_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
@@ -267,7 +264,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
response = self._workflow_node_finish_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
@@ -278,69 +275,61 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
response = self._workflow_node_finish_to_stream_response(
event=event,
task_id=self._application_generate_entity.task_id,
workflow_node_execution=workflow_node_execution
workflow_node_execution=workflow_node_execution,
)
if response:
yield response
elif isinstance(event, QueueParallelBranchRunStartedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_parallel_branch_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueParallelBranchRunSucceededEvent | QueueParallelBranchRunFailedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_parallel_branch_finished_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationStartEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_start_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationNextEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_next_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueIterationCompletedEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
yield self._workflow_iteration_completed_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run,
event=event
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run, event=event
)
elif isinstance(event, QueueWorkflowSucceededEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
if not graph_runtime_state:
raise Exception('Graph runtime state not initialized.')
raise Exception("Graph runtime state not initialized.")
workflow_run = self._handle_workflow_run_success(
workflow_run=workflow_run,
start_at=graph_runtime_state.start_at,
total_tokens=graph_runtime_state.total_tokens,
total_steps=graph_runtime_state.node_run_steps,
outputs=json.dumps(event.outputs) if isinstance(event, QueueWorkflowSucceededEvent) and event.outputs else None,
outputs=json.dumps(event.outputs)
if isinstance(event, QueueWorkflowSucceededEvent) and event.outputs
else None,
conversation_id=None,
trace_manager=trace_manager,
)
@@ -349,22 +338,23 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
self._save_workflow_app_log(workflow_run)
yield self._workflow_finish_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
elif isinstance(event, QueueWorkflowFailedEvent | QueueStopEvent):
if not workflow_run:
raise Exception('Workflow run not initialized.')
raise Exception("Workflow run not initialized.")
if not graph_runtime_state:
raise Exception('Graph runtime state not initialized.')
raise Exception("Graph runtime state not initialized.")
workflow_run = self._handle_workflow_run_failed(
workflow_run=workflow_run,
start_at=graph_runtime_state.start_at,
total_tokens=graph_runtime_state.total_tokens,
total_steps=graph_runtime_state.node_run_steps,
status=WorkflowRunStatus.FAILED if isinstance(event, QueueWorkflowFailedEvent) else WorkflowRunStatus.STOPPED,
status=WorkflowRunStatus.FAILED
if isinstance(event, QueueWorkflowFailedEvent)
else WorkflowRunStatus.STOPPED,
error=event.error if isinstance(event, QueueWorkflowFailedEvent) else event.get_stop_reason(),
conversation_id=None,
trace_manager=trace_manager,
@@ -374,8 +364,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
self._save_workflow_app_log(workflow_run)
yield self._workflow_finish_to_stream_response(
task_id=self._application_generate_entity.task_id,
workflow_run=workflow_run
task_id=self._application_generate_entity.task_id, workflow_run=workflow_run
)
elif isinstance(event, QueueTextChunkEvent):
delta_text = event.text
@@ -394,7 +383,6 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
if tts_publisher:
tts_publisher.publish(None)
def _save_workflow_app_log(self, workflow_run: WorkflowRun) -> None:
"""
Save workflow app log.
@@ -417,7 +405,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
workflow_app_log.workflow_id = workflow_run.workflow_id
workflow_app_log.workflow_run_id = workflow_run.id
workflow_app_log.created_from = created_from.value
workflow_app_log.created_by_role = 'account' if isinstance(self._user, Account) else 'end_user'
workflow_app_log.created_by_role = "account" if isinstance(self._user, Account) else "end_user"
workflow_app_log.created_by = self._user.id
db.session.add(workflow_app_log)
@@ -431,8 +419,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
:return:
"""
response = TextChunkStreamResponse(
task_id=self._application_generate_entity.task_id,
data=TextChunkStreamResponse.Data(text=text)
task_id=self._application_generate_entity.task_id, data=TextChunkStreamResponse.Data(text=text)
)
return response

154
api/core/app/apps/workflow_app_runner.py
View File

@@ -58,89 +58,86 @@ class WorkflowBasedAppRunner(AppRunner):
"""
Init graph
"""
if 'nodes' not in graph_config or 'edges' not in graph_config:
raise ValueError('nodes or edges not found in workflow graph')
if "nodes" not in graph_config or "edges" not in graph_config:
raise ValueError("nodes or edges not found in workflow graph")
if not isinstance(graph_config.get('nodes'), list):
raise ValueError('nodes in workflow graph must be a list')
if not isinstance(graph_config.get("nodes"), list):
raise ValueError("nodes in workflow graph must be a list")
if not isinstance(graph_config.get('edges'), list):
raise ValueError('edges in workflow graph must be a list')
if not isinstance(graph_config.get("edges"), list):
raise ValueError("edges in workflow graph must be a list")
# init graph
graph = Graph.init(
graph_config=graph_config
)
graph = Graph.init(graph_config=graph_config)
if not graph:
raise ValueError('graph not found in workflow')
raise ValueError("graph not found in workflow")
return graph
def _get_graph_and_variable_pool_of_single_iteration(
self,
workflow: Workflow,
node_id: str,
user_inputs: dict,
) -> tuple[Graph, VariablePool]:
self,
workflow: Workflow,
node_id: str,
user_inputs: dict,
) -> tuple[Graph, VariablePool]:
"""
Get variable pool of single iteration
"""
# fetch workflow graph
graph_config = workflow.graph_dict
if not graph_config:
raise ValueError('workflow graph not found')
raise ValueError("workflow graph not found")
graph_config = cast(dict[str, Any], graph_config)
if 'nodes' not in graph_config or 'edges' not in graph_config:
raise ValueError('nodes or edges not found in workflow graph')
if "nodes" not in graph_config or "edges" not in graph_config:
raise ValueError("nodes or edges not found in workflow graph")
if not isinstance(graph_config.get('nodes'), list):
raise ValueError('nodes in workflow graph must be a list')
if not isinstance(graph_config.get("nodes"), list):
raise ValueError("nodes in workflow graph must be a list")
if not isinstance(graph_config.get('edges'), list):
raise ValueError('edges in workflow graph must be a list')
if not isinstance(graph_config.get("edges"), list):
raise ValueError("edges in workflow graph must be a list")
# filter nodes only in iteration
node_configs = [
node for node in graph_config.get('nodes', [])
if node.get('id') == node_id or node.get('data', {}).get('iteration_id', '') == node_id
node
for node in graph_config.get("nodes", [])
if node.get("id") == node_id or node.get("data", {}).get("iteration_id", "") == node_id
]
graph_config['nodes'] = node_configs
graph_config["nodes"] = node_configs
node_ids = [node.get('id') for node in node_configs]
node_ids = [node.get("id") for node in node_configs]
# filter edges only in iteration
edge_configs = [
edge for edge in graph_config.get('edges', [])
if (edge.get('source') is None or edge.get('source') in node_ids)
and (edge.get('target') is None or edge.get('target') in node_ids)
edge
for edge in graph_config.get("edges", [])
if (edge.get("source") is None or edge.get("source") in node_ids)
and (edge.get("target") is None or edge.get("target") in node_ids)
]
graph_config['edges'] = edge_configs
graph_config["edges"] = edge_configs
# init graph
graph = Graph.init(
graph_config=graph_config,
root_node_id=node_id
)
graph = Graph.init(graph_config=graph_config, root_node_id=node_id)
if not graph:
raise ValueError('graph not found in workflow')
raise ValueError("graph not found in workflow")
# fetch node config from node id
iteration_node_config = None
for node in node_configs:
if node.get('id') == node_id:
if node.get("id") == node_id:
iteration_node_config = node
break
if not iteration_node_config:
raise ValueError('iteration node id not found in workflow graph')
raise ValueError("iteration node id not found in workflow graph")
# Get node class
node_type = NodeType.value_of(iteration_node_config.get('data', {}).get('type'))
node_type = NodeType.value_of(iteration_node_config.get("data", {}).get("type"))
node_cls = node_classes.get(node_type)
node_cls = cast(type[BaseNode], node_cls)
@@ -153,8 +150,7 @@ class WorkflowBasedAppRunner(AppRunner):
try:
variable_mapping = node_cls.extract_variable_selector_to_variable_mapping(
graph_config=workflow.graph_dict,
config=iteration_node_config
graph_config=workflow.graph_dict, config=iteration_node_config
)
except NotImplementedError:
variable_mapping = {}
@@ -165,7 +161,7 @@ class WorkflowBasedAppRunner(AppRunner):
variable_pool=variable_pool,
tenant_id=workflow.tenant_id,
node_type=node_type,
node_data=IterationNodeData(**iteration_node_config.get('data', {}))
node_data=IterationNodeData(**iteration_node_config.get("data", {})),
)
return graph, variable_pool
@@ -178,18 +174,12 @@ class WorkflowBasedAppRunner(AppRunner):
"""
if isinstance(event, GraphRunStartedEvent):
self._publish_event(
QueueWorkflowStartedEvent(
graph_runtime_state=workflow_entry.graph_engine.graph_runtime_state
)
QueueWorkflowStartedEvent(graph_runtime_state=workflow_entry.graph_engine.graph_runtime_state)
)
elif isinstance(event, GraphRunSucceededEvent):
self._publish_event(
QueueWorkflowSucceededEvent(outputs=event.outputs)
)
self._publish_event(QueueWorkflowSucceededEvent(outputs=event.outputs))
elif isinstance(event, GraphRunFailedEvent):
self._publish_event(
QueueWorkflowFailedEvent(error=event.error)
)
self._publish_event(QueueWorkflowFailedEvent(error=event.error))
elif isinstance(event, NodeRunStartedEvent):
self._publish_event(
QueueNodeStartedEvent(
@@ -204,7 +194,7 @@ class WorkflowBasedAppRunner(AppRunner):
start_at=event.route_node_state.start_at,
node_run_index=event.route_node_state.index,
predecessor_node_id=event.predecessor_node_id,
in_iteration_id=event.in_iteration_id
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, NodeRunSucceededEvent):
@@ -220,14 +210,18 @@ class WorkflowBasedAppRunner(AppRunner):
parent_parallel_start_node_id=event.parent_parallel_start_node_id,
start_at=event.route_node_state.start_at,
inputs=event.route_node_state.node_run_result.inputs
if event.route_node_state.node_run_result else {},
if event.route_node_state.node_run_result
else {},
process_data=event.route_node_state.node_run_result.process_data
if event.route_node_state.node_run_result else {},
if event.route_node_state.node_run_result
else {},
outputs=event.route_node_state.node_run_result.outputs
if event.route_node_state.node_run_result else {},
if event.route_node_state.node_run_result
else {},
execution_metadata=event.route_node_state.node_run_result.metadata
if event.route_node_state.node_run_result else {},
in_iteration_id=event.in_iteration_id
if event.route_node_state.node_run_result
else {},
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, NodeRunFailedEvent):
@@ -243,16 +237,18 @@ class WorkflowBasedAppRunner(AppRunner):
parent_parallel_start_node_id=event.parent_parallel_start_node_id,
start_at=event.route_node_state.start_at,
inputs=event.route_node_state.node_run_result.inputs
if event.route_node_state.node_run_result else {},
process_data=event.route_node_state.node_run_result.process_data
if event.route_node_state.node_run_result else {},
outputs=event.route_node_state.node_run_result.outputs
if event.route_node_state.node_run_result else {},
error=event.route_node_state.node_run_result.error
if event.route_node_state.node_run_result
and event.route_node_state.node_run_result.error
else {},
process_data=event.route_node_state.node_run_result.process_data
if event.route_node_state.node_run_result
else {},
outputs=event.route_node_state.node_run_result.outputs
if event.route_node_state.node_run_result
else {},
error=event.route_node_state.node_run_result.error
if event.route_node_state.node_run_result and event.route_node_state.node_run_result.error
else "Unknown error",
in_iteration_id=event.in_iteration_id
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, NodeRunStreamChunkEvent):
@@ -260,14 +256,13 @@ class WorkflowBasedAppRunner(AppRunner):
QueueTextChunkEvent(
text=event.chunk_content,
from_variable_selector=event.from_variable_selector,
in_iteration_id=event.in_iteration_id
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, NodeRunRetrieverResourceEvent):
self._publish_event(
QueueRetrieverResourcesEvent(
retriever_resources=event.retriever_resources,
in_iteration_id=event.in_iteration_id
retriever_resources=event.retriever_resources, in_iteration_id=event.in_iteration_id
)
)
elif isinstance(event, ParallelBranchRunStartedEvent):
@@ -277,7 +272,7 @@ class WorkflowBasedAppRunner(AppRunner):
parallel_start_node_id=event.parallel_start_node_id,
parent_parallel_id=event.parent_parallel_id,
parent_parallel_start_node_id=event.parent_parallel_start_node_id,
in_iteration_id=event.in_iteration_id
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, ParallelBranchRunSucceededEvent):
@@ -287,7 +282,7 @@ class WorkflowBasedAppRunner(AppRunner):
parallel_start_node_id=event.parallel_start_node_id,
parent_parallel_id=event.parent_parallel_id,
parent_parallel_start_node_id=event.parent_parallel_start_node_id,
in_iteration_id=event.in_iteration_id
in_iteration_id=event.in_iteration_id,
)
)
elif isinstance(event, ParallelBranchRunFailedEvent):
@@ -298,7 +293,7 @@ class WorkflowBasedAppRunner(AppRunner):
parent_parallel_id=event.parent_parallel_id,
parent_parallel_start_node_id=event.parent_parallel_start_node_id,
in_iteration_id=event.in_iteration_id,
error=event.error
error=event.error,
)
)
elif isinstance(event, IterationRunStartedEvent):
@@ -316,7 +311,7 @@ class WorkflowBasedAppRunner(AppRunner):
node_run_index=workflow_entry.graph_engine.graph_runtime_state.node_run_steps,
inputs=event.inputs,
predecessor_node_id=event.predecessor_node_id,
metadata=event.metadata
metadata=event.metadata,
)
)
elif isinstance(event, IterationRunNextEvent):
@@ -352,7 +347,7 @@ class WorkflowBasedAppRunner(AppRunner):
outputs=event.outputs,
metadata=event.metadata,
steps=event.steps,
error=event.error if isinstance(event, IterationRunFailedEvent) else None
error=event.error if isinstance(event, IterationRunFailedEvent) else None,
)
)
@@ -371,9 +366,6 @@ class WorkflowBasedAppRunner(AppRunner):
# return workflow
return workflow
def _publish_event(self, event: AppQueueEvent) -> None:
self.queue_manager.publish(
event,
PublishFrom.APPLICATION_MANAGER
)
self.queue_manager.publish(event, PublishFrom.APPLICATION_MANAGER)

198
api/core/app/apps/workflow_logging_callback.py
View File

@@ -30,169 +30,145 @@ _TEXT_COLOR_MAPPING = {
class WorkflowLoggingCallback(WorkflowCallback):
def __init__(self) -> None:
self.current_node_id = None
def on_event(
self,
event: GraphEngineEvent
) -> None:
def on_event(self, event: GraphEngineEvent) -> None:
if isinstance(event, GraphRunStartedEvent):
self.print_text("\n[GraphRunStartedEvent]", color='pink')
self.print_text("\n[GraphRunStartedEvent]", color="pink")
elif isinstance(event, GraphRunSucceededEvent):
self.print_text("\n[GraphRunSucceededEvent]", color='green')
self.print_text("\n[GraphRunSucceededEvent]", color="green")
elif isinstance(event, GraphRunFailedEvent):
self.print_text(f"\n[GraphRunFailedEvent] reason: {event.error}", color='red')
self.print_text(f"\n[GraphRunFailedEvent] reason: {event.error}", color="red")
elif isinstance(event, NodeRunStartedEvent):
self.on_workflow_node_execute_started(
event=event
)
self.on_workflow_node_execute_started(event=event)
elif isinstance(event, NodeRunSucceededEvent):
self.on_workflow_node_execute_succeeded(
event=event
)
self.on_workflow_node_execute_succeeded(event=event)
elif isinstance(event, NodeRunFailedEvent):
self.on_workflow_node_execute_failed(
event=event
)
self.on_workflow_node_execute_failed(event=event)
elif isinstance(event, NodeRunStreamChunkEvent):
self.on_node_text_chunk(
event=event
)
self.on_node_text_chunk(event=event)
elif isinstance(event, ParallelBranchRunStartedEvent):
self.on_workflow_parallel_started(
event=event
)
self.on_workflow_parallel_started(event=event)
elif isinstance(event, ParallelBranchRunSucceededEvent | ParallelBranchRunFailedEvent):
self.on_workflow_parallel_completed(
event=event
)
self.on_workflow_parallel_completed(event=event)
elif isinstance(event, IterationRunStartedEvent):
self.on_workflow_iteration_started(
event=event
)
self.on_workflow_iteration_started(event=event)
elif isinstance(event, IterationRunNextEvent):
self.on_workflow_iteration_next(
event=event
)
self.on_workflow_iteration_next(event=event)
elif isinstance(event, IterationRunSucceededEvent | IterationRunFailedEvent):
self.on_workflow_iteration_completed(
event=event
)
self.on_workflow_iteration_completed(event=event)
else:
self.print_text(f"\n[{event.__class__.__name__}]", color='blue')
self.print_text(f"\n[{event.__class__.__name__}]", color="blue")
def on_workflow_node_execute_started(
self,
event: NodeRunStartedEvent
) -> None:
def on_workflow_node_execute_started(self, event: NodeRunStartedEvent) -> None:
"""
Workflow node execute started
"""
self.print_text("\n[NodeRunStartedEvent]", color='yellow')
self.print_text(f"Node ID: {event.node_id}", color='yellow')
self.print_text(f"Node Title: {event.node_data.title}", color='yellow')
self.print_text(f"Type: {event.node_type.value}", color='yellow')
self.print_text("\n[NodeRunStartedEvent]", color="yellow")
self.print_text(f"Node ID: {event.node_id}", color="yellow")
self.print_text(f"Node Title: {event.node_data.title}", color="yellow")
self.print_text(f"Type: {event.node_type.value}", color="yellow")
def on_workflow_node_execute_succeeded(
self,
event: NodeRunSucceededEvent
) -> None:
def on_workflow_node_execute_succeeded(self, event: NodeRunSucceededEvent) -> None:
"""
Workflow node execute succeeded
"""
route_node_state = event.route_node_state
self.print_text("\n[NodeRunSucceededEvent]", color='green')
self.print_text(f"Node ID: {event.node_id}", color='green')
self.print_text(f"Node Title: {event.node_data.title}", color='green')
self.print_text(f"Type: {event.node_type.value}", color='green')
self.print_text("\n[NodeRunSucceededEvent]", color="green")
self.print_text(f"Node ID: {event.node_id}", color="green")
self.print_text(f"Node Title: {event.node_data.title}", color="green")
self.print_text(f"Type: {event.node_type.value}", color="green")
if route_node_state.node_run_result:
node_run_result = route_node_state.node_run_result
self.print_text(f"Inputs: {jsonable_encoder(node_run_result.inputs) if node_run_result.inputs else ''}",
color='green')
self.print_text(
f"Inputs: {jsonable_encoder(node_run_result.inputs) if node_run_result.inputs else ''}", color="green"
)
self.print_text(
f"Process Data: {jsonable_encoder(node_run_result.process_data) if node_run_result.process_data else ''}",
color='green')
self.print_text(f"Outputs: {jsonable_encoder(node_run_result.outputs) if node_run_result.outputs else ''}",
color='green')
color="green",
)
self.print_text(
f"Outputs: {jsonable_encoder(node_run_result.outputs) if node_run_result.outputs else ''}",
color="green",
)
self.print_text(
f"Metadata: {jsonable_encoder(node_run_result.metadata) if node_run_result.metadata else ''}",
color='green')
color="green",
)
def on_workflow_node_execute_failed(
self,
event: NodeRunFailedEvent
) -> None:
def on_workflow_node_execute_failed(self, event: NodeRunFailedEvent) -> None:
"""
Workflow node execute failed
"""
route_node_state = event.route_node_state
self.print_text("\n[NodeRunFailedEvent]", color='red')
self.print_text(f"Node ID: {event.node_id}", color='red')
self.print_text(f"Node Title: {event.node_data.title}", color='red')
self.print_text(f"Type: {event.node_type.value}", color='red')
self.print_text("\n[NodeRunFailedEvent]", color="red")
self.print_text(f"Node ID: {event.node_id}", color="red")
self.print_text(f"Node Title: {event.node_data.title}", color="red")
self.print_text(f"Type: {event.node_type.value}", color="red")
if route_node_state.node_run_result:
node_run_result = route_node_state.node_run_result
self.print_text(f"Error: {node_run_result.error}", color='red')
self.print_text(f"Inputs: {jsonable_encoder(node_run_result.inputs) if node_run_result.inputs else ''}",
color='red')
self.print_text(f"Error: {node_run_result.error}", color="red")
self.print_text(
f"Inputs: {jsonable_encoder(node_run_result.inputs) if node_run_result.inputs else ''}", color="red"
)
self.print_text(
f"Process Data: {jsonable_encoder(node_run_result.process_data) if node_run_result.process_data else ''}",
color='red')
self.print_text(f"Outputs: {jsonable_encoder(node_run_result.outputs) if node_run_result.outputs else ''}",
color='red')
color="red",
)
self.print_text(
f"Outputs: {jsonable_encoder(node_run_result.outputs) if node_run_result.outputs else ''}", color="red"
)
def on_node_text_chunk(
self,
event: NodeRunStreamChunkEvent
) -> None:
def on_node_text_chunk(self, event: NodeRunStreamChunkEvent) -> None:
"""
Publish text chunk
"""
route_node_state = event.route_node_state
if not self.current_node_id or self.current_node_id != route_node_state.node_id:
self.current_node_id = route_node_state.node_id
self.print_text('\n[NodeRunStreamChunkEvent]')
self.print_text("\n[NodeRunStreamChunkEvent]")
self.print_text(f"Node ID: {route_node_state.node_id}")
node_run_result = route_node_state.node_run_result
if node_run_result:
self.print_text(
f"Metadata: {jsonable_encoder(node_run_result.metadata) if node_run_result.metadata else ''}")
f"Metadata: {jsonable_encoder(node_run_result.metadata) if node_run_result.metadata else ''}"
)
self.print_text(event.chunk_content, color="pink", end="")
def on_workflow_parallel_started(
self,
event: ParallelBranchRunStartedEvent
) -> None:
def on_workflow_parallel_started(self, event: ParallelBranchRunStartedEvent) -> None:
"""
Publish parallel started
"""
self.print_text("\n[ParallelBranchRunStartedEvent]", color='blue')
self.print_text(f"Parallel ID: {event.parallel_id}", color='blue')
self.print_text(f"Branch ID: {event.parallel_start_node_id}", color='blue')
self.print_text("\n[ParallelBranchRunStartedEvent]", color="blue")
self.print_text(f"Parallel ID: {event.parallel_id}", color="blue")
self.print_text(f"Branch ID: {event.parallel_start_node_id}", color="blue")
if event.in_iteration_id:
self.print_text(f"Iteration ID: {event.in_iteration_id}", color='blue')
self.print_text(f"Iteration ID: {event.in_iteration_id}", color="blue")
def on_workflow_parallel_completed(
self,
event: ParallelBranchRunSucceededEvent | ParallelBranchRunFailedEvent
self, event: ParallelBranchRunSucceededEvent | ParallelBranchRunFailedEvent
) -> None:
"""
Publish parallel completed
"""
if isinstance(event, ParallelBranchRunSucceededEvent):
color = 'blue'
color = "blue"
elif isinstance(event, ParallelBranchRunFailedEvent):
color = 'red'
color = "red"
self.print_text("\n[ParallelBranchRunSucceededEvent]" if isinstance(event, ParallelBranchRunSucceededEvent) else "\n[ParallelBranchRunFailedEvent]", color=color)
self.print_text(
"\n[ParallelBranchRunSucceededEvent]"
if isinstance(event, ParallelBranchRunSucceededEvent)
else "\n[ParallelBranchRunFailedEvent]",
color=color,
)
self.print_text(f"Parallel ID: {event.parallel_id}", color=color)
self.print_text(f"Branch ID: {event.parallel_start_node_id}", color=color)
if event.in_iteration_id:
@@ -201,43 +177,37 @@ class WorkflowLoggingCallback(WorkflowCallback):
if isinstance(event, ParallelBranchRunFailedEvent):
self.print_text(f"Error: {event.error}", color=color)
def on_workflow_iteration_started(
self,
event: IterationRunStartedEvent
) -> None:
def on_workflow_iteration_started(self, event: IterationRunStartedEvent) -> None:
"""
Publish iteration started
"""
self.print_text("\n[IterationRunStartedEvent]", color='blue')
self.print_text(f"Iteration Node ID: {event.iteration_id}", color='blue')
self.print_text("\n[IterationRunStartedEvent]", color="blue")
self.print_text(f"Iteration Node ID: {event.iteration_id}", color="blue")
def on_workflow_iteration_next(
self,
event: IterationRunNextEvent
) -> None:
def on_workflow_iteration_next(self, event: IterationRunNextEvent) -> None:
"""
Publish iteration next
"""
self.print_text("\n[IterationRunNextEvent]", color='blue')
self.print_text(f"Iteration Node ID: {event.iteration_id}", color='blue')
self.print_text(f"Iteration Index: {event.index}", color='blue')
self.print_text("\n[IterationRunNextEvent]", color="blue")
self.print_text(f"Iteration Node ID: {event.iteration_id}", color="blue")
self.print_text(f"Iteration Index: {event.index}", color="blue")
def on_workflow_iteration_completed(
self,
event: IterationRunSucceededEvent | IterationRunFailedEvent
) -> None:
def on_workflow_iteration_completed(self, event: IterationRunSucceededEvent | IterationRunFailedEvent) -> None:
"""
Publish iteration completed
"""
self.print_text("\n[IterationRunSucceededEvent]" if isinstance(event, IterationRunSucceededEvent) else "\n[IterationRunFailedEvent]", color='blue')
self.print_text(f"Node ID: {event.iteration_id}", color='blue')
self.print_text(
"\n[IterationRunSucceededEvent]"
if isinstance(event, IterationRunSucceededEvent)
else "\n[IterationRunFailedEvent]",
color="blue",
)
self.print_text(f"Node ID: {event.iteration_id}", color="blue")
def print_text(
self, text: str, color: Optional[str] = None, end: str = "\n"
) -> None:
def print_text(self, text: str, color: Optional[str] = None, end: str = "\n") -> None:
"""Print text with highlighting and no end characters."""
text_to_print = self._get_colored_text(text, color) if color else text
print(f'{text_to_print}', end=end)
print(f"{text_to_print}", end=end)
def _get_colored_text(self, text: str, color: str) -> str:
"""Get colored text."""